The Wisdom of Crowds
March 26th, 2006
Just finished listening to the audio of James Surowiecki’s presentation at SXSW. I’ve not yet read James’ book, but the gist of it is that “under the right circumstances, groups are remarkably intelligent, and are often smarter than the smartest people in them.”
The key seems to be ‘under the right circumstances‘. Surowiecki’s conditions for the right circumstances are diversity and independance of opinion and a decentralization of knowledge. Then there must be a means of aggregating these individual views to form a collective decision. Further, to make use of this knowledge, there must be an actual ‘right’ or ‘wrong’ answer and a clear understanding as to what the problem is.
While there may be evidence in the book that contradicts me on this, I would say that the crowd must have some inital basis of facts and experience from which to work from.
James gives an example of naval officer John Craven who was searching for a missing submarine. Craven took the best guess of a wide number of professionals (mathematicians, salvagers, naval officers, submarine specialists). Not one of these expert guesses were anywhere near the sub. However crunching the individual gueses together using a statistical formula, Craven was able to predict within a couple hundred meters the actual position of the missing vessel. This is a very powerful case for the wisdom of arguments but… would Craven have been successful if his group had consisted of wheat farmers, beaver trappers and vacuum cleaner repairmen?
As Rob May points out in an essay critiquing the Wisdom of Crowds, “the crowd would say that evolution is false. Does that matter? No, because most of the crowd doesn’t know the difference between a genotype and a phenotype. If you don’t know basic biology then your opinion on evolution is irrelevant.” Rob makes some other persuasive points worth a read over at his blog Business Pundit.
I’m not persuaded that the crowd opinion, aggregated or otherwise, holds wisdom, I will grant that there is distributed knowledge and that James’ conditions for wise crowds provides an excellent means of extracting that knowledge with minimal bias or outside influence.
