2 min read

What Crowd Wisdom Means for Data

In The Wisdom of Crowds, journalist James Surowiecki theorizes information collected and aggregated from groups oftentimes provides better results than those provided by single group members or experts. An oft-cited anecdote from the book relates Sir Francis Galton's observations at a 1906 livestock fair. As part of a contest, fairgoers were invited to guess the weight of a butchered ox. Although hundreds of contestants failed to guess the real weight of 1,198 pounds, the mean of the individual guesses rang in at 1,197 pounds.

Today this statistical phenomenon goes by the name crowd wisdom and is defined as “a diverse collection of independently-deciding individuals whose individual biases cancel each other out.” With a business that relies on the accurate collection, aggregation, and analysis of market price data, this definition requires we ask, “If taking the average of individual opinions negates bias, then what is the problem with survey data?”
 
In part, the key lies with the word diversity. A wise crowd must be diverse, independent, and decentralized. Surowiecki notes, “The problem is that groups are only smart when the people in them are as independent as possible. This is the paradox of the wisdom of crowds.”
 
A Wired magazine article describes this paradox in terms of a study where crowd wisdom was negated when participants knew their peers’ bias. The average answers of independent test subjects became more accurate as testing progressed, but the answers provided by those test subjects who were told what their peers were guessing became less accurate. This trend reversal was attributed to social influences that made opinions less diverse, clustered correct answers at the periphery, and caused participants to become more confident about their responses.
 
Researchers explained, “Although groups are initially wise, knowledge about estimates of others narrows the diversity of opinions to such an extent that it undermines collective wisdom. Even mild social influence can undermine the wisdom of crowd effect.”
This effect is intensified in market systems that rely on collective assessment. “Opinion polls…largely promote information feedback and therefore trigger convergence of how we judge the facts. The wisdom of crowds is valuable, but used improperly it creates overconfidence in possibly false beliefs.”
 
Market data unearthed by survey is sourced from a small sample of non-diverse players and collected by word of mouth where a two-way conversation is taking place about the market. Furthermore, the “crowd” is no crowd at all; it is generally comprised of a small group of insiders from the same industry segment (buyers or sellers, not both). In short, the sample is not large enough and the biases are not diverse enough for an accurate reading of market price to result.
 
Forest2Market, from its very inception, has pledged to collect only transaction-level data free from survey bias. In addition, our substantial market share (as high as 90 percent in some geographies and industries) provides us with the necessary crowd to ensure accuracy. The only bias you will find at Forest2Market is that we will always be an advocate for working forests.
economic outlook

Comments

Dharmendra Daukia

09-02-2014

I have read the Book of James Surowiecki. Heard his audios from youtube.

I am convinced about each and every concept taken in WHOLISTIC manner.

I have done forecasting of pulp wood from agro-forestry in three distinct catchments in India in three years - 2004, 2005, 2006. The results were not believed by any one at first but 2-3 years down the line they were 80% accurate when crops came to market.

One may call these as opinion poll or any other jargon, but the methodology taken in totality gives very accurate results.

One has to set the stage as outlined in the book - large number, diverse, uni-biased, not talking to each other etc etc.