Wisdom of Groups

If you have a specific problem you need to solve, ask several people who have a vested interest and are not just picked randomly off the street

In his splendid book, ‘Wisdom of Crowds’, subtitled ‘Why the many are smarter than the few, James Sorowiecki kicks off by telling the tale of Francis Galton, an English scientist who in 1906 headed for a country fair where local farmers gathered, not least to appraise the quality of each other’s cattle, sheep, chickens, horses and pigs

Whilst there, Galton chanced upon a weight-judging competition:

  • A fat ox was on display
  • A gathering crowd were placing wagers on what the weight of the ox would be after it had been ‘slaughtered and dressed’
  • The best guesses would receive prizes
  • 800 tried their luck – they were a diverse lot – many were butchers and farmers but quite a few were ‘non-experts’, having no insider-knowledge
  • Afterwards, Galton got hold of their guesses and calculated their mean value – the collective wisdom of the participants
  • The mean guess was 1,197 pounds – the actual weight of the ox was 1,198 pounds

 

Galton wrote later: “The result seems more creditable to the trustworthiness of a democratic judgement than might have been expected” – (It certainly gives one more faith in the result of general elections given the diversity of the electorate)

He concluded: “If one employs the collective judgement of people in the know, and asks them to use their experience, common-sense, intuition and gut-feel, one can get a good feel for important relationships”

We used a similar approach – aka the Delphi Technique – at British Steel when trying to establish how steel would be made 10 years ahead

The BS board of directors believed that, whilst unknown external innovations could have a big impact, much of what would shape the steel sector 10 years on was probably known already by internal specialists – hence, a sample of them were selected and asked what they thought might happen:

  • First, individually and independently
  • Some forecasts were pretty wild e.g. ‘Steel will be produced on the moon, using the iron ore there;’
  • Then a report on their first views was written and posted to each of them – they were asked to read it and reconsider
  • Their second views converged, made good sense and were thought realistic
  • British Steel thus used them in their corporate plan that year – the confidence shown in them at the time was well justified some 10 years later
  • (for details, see Section 6.6.5 in the new version of our book Productivity Knowhow)

 

Contrast the above approach with Brainstorming, a technique much promoted by many:

  • Brainstorming claims to encourage creative thinking by allowing free-wheeling of a group’s thoughts
  • It’s said to be a formalised way of producing dozens of ideas to solve real problems
  • However, it does not require participants to be ‘in the know’
  • And it only allows them a short time to generate ideas – usually under 30 minutes
  • Hence, it usually fails because of lack of time to think plus pressure and embarrassment rife among individuals in the group which stunts thinking and creativity
  • Hence, most ideas generated are usually daft and/ or unusable

 

But now, Tom Whipple, science editor of The Times, reports on a new way to trawl ideas for solving specific problems – he starts:

  • “You have 100 people in front of you and need to know three things:
    • The height of Mount Everest?
    • The weight of a cow?
    • The volume of the Albert Hall?
  • Would you ask each of them individually and average the results, or let them come to a consensus?

 

Apparently, according to a new study, you should do a mix of the two – Whipple claims researchers have just shown that the wisdom of a crowd is not that wise after all – but neither is groupthink (albeit without proof for either claim):

  • Wisdom of a crowd requires all in the crowd to have ‘a bit of expertise’ about the subject involved – and when their results are averaged, outlying views are thought to nullify each other (which, overall, is unlikely)
  • Groupthink has a bad reputation because of a desire for conformity and belonging which leads a group to converge on extreme ideas without challenging them

 

However, by taking the best of both and splitting the 100 people into 20 groups of five, the claim is made that you produce vastly superior results

Scientists have thus tried to find this happy medium and assessed the ‘wisdom of groups’

  • They asked a sample of 5,000 people a set of simple questions such as ‘the height of the Eiffel Tower’
  • Each person answered individually – again, the average converged on the right answer
  • Then they asked the same questions, but people deliberated in groups of five first to produce a consensus answer
  • When these results were averaged  they found a dramatic difference
  • The average of four collective estimates was more accurate than the average of 1,400 initial estimates

 

So there you have it – the wisdom of groups from within crowds should win the day – or bet

Leave a Reply

Your e-mail address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.