About Rationally Speaking
Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.
Thursday, September 29, 2005
Are crowds wise?
Almost eight hundred people tried, many of whom were not experts at judging the weight of large mammals. Galton collected the data, wishing to test the hypothesis that the average would be significantly off the mark, and that a few people -- presumably with more expertise than the average -- would instead get very close to the real number. He published the results in the journal Nature, where he noted his surprise at the fact that the weight of the ox, once slaughtered and dressed (that was the bet), was 1,197 pounds. The actual result was 1,198, an error rate of only 0.0008! Galton's comment was that "the result seems more creditable to the trustworthiness of a democratic judgment than might have been expected."
As Surowiecki notes, this sort of phenomenon -- to which he refers to as the "wisdom" of the crowds -- has been repeatedly documented since, and the book presents several intriguing examples. The thing is a bit puzzling, until one realizes that it works under certain sets of circumstances and for certain types of problems.
The conditions include: diversity of opinion (obviously), independence (i.e., people are not communicating and/or arguing with each other), decentralization (people can draw on local knowledge, including the possibility of specialization), and aggregation (there has to be a mechanism to turn independent private judgments into a collective decision).
This works because of what really is a simple truth in mathematical statistics: the errors made by each independent individual will be of similar magnitude but different direction and, if the group is large enough, on average they will cancel each other, so that the group's estimate will turn out to be remarkably accurate.
Of course, this works for certain categories of problems only, though Surowiecki attempts to show that these categories are so broad that the method should be used to solve all sorts of human problems (which is, of course, a gross exaggeration). The kinds of problems where crowds might be "wiser" than most individuals include: cognition problems (such as the problem of the ox mentioned above), coordination problems (where coordination with other members of a group is required, as in walking a crowded street), and cooperation problems (where one wants self-interested people to cooperate, as in paying taxes). While these categories seem indeed almost all-encompassing, Surowiecki himself gives plenty of examples where the crowds don't perform better than the experts, because often one or more of the conditions given above are violated (for example, faculty meetings tend to be particularly inefficient, because they are small groups, opinions are not independent, and they tend not to be sufficiently diverse). Moreover, it isn't really all that clear that the "wisdom" involved in the three cases is really the same sort of phenomenon (after all, the three kinds of tasks are very different, and presumably they depend on different cognitive abilities).
This makes for very interesting reading, but there are two major caveats that need to be kept in mind. First, none of the three categories of problems identified by Surowiecki has anything to do with "wisdom." Which undermines one of the major messages of the book, that democracies are a "wise" form of government because they are the expression of the will of a large number of people. Second, there is a more subtle, almost anti-intellectual, message here, that "experts" aren't to be trusted, because they don't do as well as crowds. This populist attitude feeds into the already generally anti-intellectual American public, and may have serious consequences when it comes to addressing problems that don't fit the classification or meet the requirements laid out by Surowiecki. Which, I suspect, happens a lot more often than the author would like to admit.