The art of spurious opinion polls

by Stephen Tall on February 20, 2007

I’m a self-confessed opinion poll-geek, as regular readers may have garnered. But I have had my fill of ‘what if’ polls making psephologically spurious claims to predict what might happen if/when Gordon Brown faces up to David Cameron and Ming Campbell.

The latest (I’m sure you’ve seen) dominates today’s Grauniad front page – the headline screams, Brown v Cameron – exclusive poll puts Labour 13 points adrift. The question which prompts this hyperbole asks how the public might vote if Mr Brown led Labour, the Tories are led by Mr Cameron, and Sir Menzies leads the Lib Dems.

Astute readers will have noticed a small flaw: one of these people currently does not lead his party. What ICM are in fact asking is: how well do you think Gordon Brown might do a job he’s not currently doing? To read anything into the answer, other than that the Chancellor currently suffers a PR deficit, is plain silly.

Opinion pollsters often defend their claim to represent a science by pointing out (quite rightly) that all they can do is take a snapshot of public opinion; that they can’t predict the future. But they do their art no good by asking questions which seek to anticipate the future. Especially when they know such polling will be spun by newspapers with their own agenda.