Watch CBSN Live

Crowdsourcing: How to Make a "Crowd" Smarter

You've heard of the wisdom of crowds, which relies on the somewhat counterintuitive fact that a group of random people can generally come up with more accurate estimates than any single expert.

But how wise is any particular crowd? And how can we make it smarter?

New research gives us a hint: It suggests that the personality of individuals in the crowd has a big effect on how 'wise' the crowd actually is.

INSEAD's Kriti Jain, J. Neil Bearden, and Allan Filipowicz started their research into the wisdom of crowds by asking 350 people take personality tests. They then grouped people into pairs. In each pair, the two people either had very similar or very different personalities. Each person was asked to predict the fate of a particular World Cup team and to estimate the number of M&Ms in a jar.

  • Pairs of people with dramatically different personalities made the most accurate predictions. Asked to predict the fortunes of a particular World Cup team, they were correct about 42% of the time. Asked to estimate the number of M&Ms, they only missed by about 23%.
  • Pairs of people with similar personalities didn't do as well. On the World Cup test, those pairs were correct about 32% of the time. They missed the number of M&Ms in the jar by about 26%.
  • Individuals did the worst. They were right on the World Cup test just under 30% of the time. They totally blew the M&M test, missing the proper number by an average of 48%.
The conclusion?
  • Having access to diverse information helps a lot. Crowdsourcing works best when members of the crowd are getting lots of different data. That causes their errors to cancel each other out.
  • Having a crowd with different personalities also improves accuracy. The researchers figured that people of very different personalities would likely rely on different information when making their estimates or predictions.
Trying this at home
You don't have to be part of a crowd to take advantage of this phenomenon: You can get a better forecast just by averaging two predictions from the same person who has been encouraged to think about something from multiple perspectives. (This process has become known as --brace yourself--dialectical bootstrapping.) Asking the same person to make estimates at different points in time also results in better accuracy.

How else might we increase the wisdom of crowds? And is there any way for the lone expert to keep up?


Photo courtesy of flickr user Amanda M Hatfield
Kimberly Weisul is a freelance writer, editor and consultant. Follow her on twitter at www.