10.10 New methods – The power of wisdom: The crowd’s ability to predict things
Martin Boon, Director, ICM
- We have a duty to challenge boundaries and come up with new ideas. Never apologize for coming up with new idea regardless of whether they work.
- Under the right circumstances, groups can be smarter than the smartest people. The average guess of people at a fair guessing the weight of ox, was exactly correct.
- For a crowd to be right
- There must be diverse opinion where each person has some information even if it’s a strange interpretation of the information
- Views are independent, not determined of the view of anyone around them
- People can use their own specialized knowledge and experience
- Some method of aggregation, for turning private into a collective number
- How has UK polling performed since 1992? 1992 wrong, 1997 wrong, 2001 right, 2005 right, 2010 ok.
- ICM added two questions to their prediction poll – “For a bit of fun, tell me what % share of the vote you think …party…. will win in the forthcoming general election?”
- Wisdom prompted predictions were the most accurate results. Wisdom outperformed conventional polling. Every poll overstated the Liberal Democrat share but the crowd wisdom got it exactly right.
- Polls find it hard to predict parties that people know are unpopular to vote for – embarrassment to say you are voting for that party.
- Wisdom approach produces a consistent share of vote compared to actual.
- But in a Wales referendum and an AV referendum, the method didn’t work as well with 9% error. Was the diversity of opinion condition present, and were they suitable informed to be able to take a smart view? Most people said they didn’t have enough information to vote well. Ensure responders know the key features of products and services if you’re going to try this.
- Wisdom predicted 18 gold medals at the Olympic games but then they got 29. You just can’t predict how well athletes are going to perform.
- Our job as researchers is to test things to destruction [i like that!]
- Wisdom could be good for market sizing and product supply estimates giving marketers an advantage ahead of product releases
- The 6 Worst Market Research Mistakes #MRX (lovestats.wordpress.com)
- 4 rules to achieve sentiment validity scores of 90% or greater #MRX (lovestats.wordpress.com)
- The Social Media Research Love Hate Debate #MRX (lovestats.wordpress.com)
- In defense of research participants #MRX (lovestats.wordpress.com)
- The loyalty card misconception #MRX (lovestats.wordpress.com)