You know something they don't know
by Duncan O'Leary
I’m enjoying the new Politicshome – it’s found a niche in providing cross-party, up the minute information à la Bloomberg. What i’m not so sure about is the feature of the website that seems to be getting all the attention – the PH100 polls that are done. (e.g. who will be the next mayor)
According to the website, the Phi100 panel is an expert panel of over 100 of the top political brains in the UK. It includes senior members of the government and some of the biggest names in political journalism. There’s a basic wisdom of crowds issue here: the best aggregated predictions come from groups where people make independent decisions and where people bring different information to answer a question.
So, for all the talented people involved, isn’t there the danger of group-think in the PH100? How much different information does each panellist bring? And how much have they all discussed the answer already, or read each other's columns? I don’t know the answer, but for the moment i’ll be sticking with who the public thinks is going to win rather than with the experts.
Sam Jones
Taking Jack's point a little further, it also brings into question scales of value, particularly where criticism is concerned.The critic, Mark Lawson, reflected in the Guardian that he had once reviewed a book by a previously unheard of American writer. In his words 'It was a matter of fine judgment whether the content (conspiracy theory) or the prose (making a crisp packet read like a sonnet in comparison) was more preposterous. I worried, though, about being too brutal to a new talent'. The book turned out to be The da Vinci Code, and it went on to sell millions.What's really interesting about this isn't just market success v. critical acclaim, it is, as Jack asks, about a clash of values (individual interest and popular acclaim v. aesthetics and critical judgement) and the balance of expertise in relation to personal choice.
Mark Hendry
Duncan is right. It is group think. These 100 people are not independent thinkers who never come into contact with each other. Anyone who watches a lot of politics shows on TV or listens to politics on the radio, or reads political articles in newspapers and magazines will quickly realise that they feature the same people over and over again, expressing their well-rehearsed opinions. Often, a commentator will write an article, which he then discusses on the radio, and then on TV. In other words, one thought gets maximum play. All of these commentators know each other and appear on the same shows, and comment on each others' work. They're not "experts"...they are people who have been deemed acceptable within a particular framework. You won't find communists or BNP people featuring amongst the "experts" because such people have opinions that are deemed too extreme. So, all you get exposed to is a very narrow range of opinions from a liberal democratic capitalist elite who relentlesly scratch each others' backs to maintain their positions. There is no "expert thought"...just the tired opinions of political hacks who are enjoying a very comfortable lifestyle, thank you very much.
Philip Conway
Mark Hendry:You're right that they aren't representative of the world as a whole, they are representative of the ruling elites, but thats kinda the point isn't it? This panel has been set up to generate predictions regarding these elites and so are well positioned to do so.
Jack Stilgoe
Isn't group-think different from this? Doesn't group-think rely on some sort of deliberation? It's all about how discussions within small groups get framed and constrained so that people forget about or feel unable to ask important questions, like with the launch decision behind Challenger, to use Surowiecki's example. The non-deliberative idea may be appropriate for elections, in which people are likely to make decisions without bashing about the pros and cons in great detail. But the big methodological issue is whether the views of experts, even in isolation, reflect those of lay people.