Thursday, April 26, 2007

Experts: Pro vs. Con

[O]ne study found that on average when doctors were 88% confident that their patient had pneumonia, in fact only 20% of such patients had pneumonia. And overconfidence is fatal in primary care.
This is from a blog post by Robin Hanson, who also describes a study showing that highly paid primary care doctors are no better at treating patients than nurse practitioners.

GlobalGiving is based to a significant degree on the "wisdom of the crowds." This is the idea that regular people, if their judgements are pooled, can often make decisions that are equal to or better than experts. For this to hold true, people have to make decisions somewhat independently, and they have to bring a diversity of perspectives and viewpoints.

If you think about it, regular economic markets are in essence a reflection of the wisdom of crowds. For all their shortcomings (and occasional need for regulation), markets still work a heck of a lot better than systems such as central planning where a few experts decide what gets produced and to whom it gets distributed.

But to be fair, the wisdom of crowds does not exclude experts. Experts are part of the crowd, too - often an important part.

Here is a nice (long) post by Larry Sanger on Edge.org that tries to bridge this issue. Larry is a co-founder of Wikipedia who has left to found Citizendium, or the Citizens' Compendium. Larry argues that delegating special authority to experts in creating and cataloguing knowledge makes sense WHEN it is accompanied by open-source peer review and input by regular people in the crowd. Wikipedia fails, he argues, by not according experts special status. Encyclopedia Brittanica fails, he says, by not allowing regular people to comment on or contribute to articles written by experts.

Over the coming months, we will be phasing in a number of new community features on GlobalGiving, including blogs and reputation mechanisms. We are currently grappling with the question of how experts will be allowed to emerge on the system and what special rights and privileges they might have. One thing you can be sure of, though: we won't recognize experts solely by their academic degrees or by where they have worked. Experts will have to demonstrate their value in the community rather than rely on paper credentials.