Monday, May 23, 2011

The crowds are not always wise!

James Surowiecki's best selling book, The Wisdom of Crowds, popularized the belief that the collective wisdom of a group of people was superior to the individual wisdom of even experts. It has generated considerable interest in the design of systems that seek to channelize the knowledge of large groups of people to say, predict events and prices. See Justin Wolfers' paper on prediction markets here.

It is based on the statistical phenomenon by which individual biases cancel each other out, distilling hundreds or thousands of individual guesses into uncannily accurate average answers. However, it assumes that the members of the crowd have a variety of opinions, and arrive at those opinions independently.

A new study of this phenomenon by Jan Lorenz and Heiko Rahut finds that contrary to conventional wisdom, groups insights could go awry if participants were influenced by the guesses of their peer group. They found that though groups are initially wise, "knowledge about estimates of others narrows the diversity of opinions to such an extent that it undermines” collective wisdom". Moreover, they found that "even mild social influence can undermine the wisdom of crowd effect". In this context, as the Wired article points out, computer modeling of crowd behavior also hints at dynamics underlying crowd breakdowns, with the balance between information flow and diverse opinions becoming skewed.

The authors recruited 144 students from ETH Zurich, made them sit in isolated cubicles and asked them to guess various indicators like Switzerland’s population density, the length of its border with Italy, the number of new immigrants to Zurich and how many crimes were committed in 2006.

At the end of each round of questioning, they were given small payments for coming close to the actual answer (signified by the gray bar). At left is the range of responses among participants who received no information about others. The findings of the study participants who were asked how many murders occurred in Switzerland in 2006 is shown in the graphic below.

The Wired article concludes,

"As testing progressed, the average answers of independent test subjects became more accurate, in keeping with the wisdom-of-crowds phenomenon. Socially influenced test subjects, however, actually became less accurate. The researchers attributed this to three effects. The first they called "social influence": Opinions became less diverse. The second effect was "range reduction": In mathematical terms, correct answers became clustered at the group’s edges. Exacerbating it all was the "confidence effect", in which students became more certain about their guesses."

As the authors claim, such false beliefs are commonplace in society, politics and markets. The herd behaviour of investors in financial markets is driven by excessive confidence generated by social influences. Opinion polls and the mass media largely promote information feedback and therefore trigger convergence of how we judge the facts and potentially create overconfidence in possibly false beliefs. Social fads and beliefs, some of which are of questionable value, become popular for no apparent reason.

In all these areas - markets, society, and politics - there are people and groups with an interest in influencing the beliefs of participants. They are vulnerable to being manipulated to suit the requirements of these vested interests. Such dissonances constitute failures in markets, politics and society.

No comments: