Festival of Nature 2014

Update on the Guess the Weight Competition

Picture of toy chimps used in the 'guess the weight' competition

Thank you to all the visitors to our stand on the weekend of the 14th and 15th June and we hope that you all had as much fun at the event as we did.

We have now weighed the chimps and their total weight came to 996 grams. Although no one got it spot on, we had 7 contestants who all guessed to within 4 grams of the correct answer. Those names went into a hat and we drew out the 3 lucky winners. All winners have now been informed, so if you haven't heard from us I'm sorry to say you haven't been lucky on this occasion.

We got these adorable guys from the Care for the Wild website, so you can still buy a chimp of your own and help support endangered wildlife: http://www.careforthewild.com/online-shop/category/1/soft-toys

Now for the science bit! We were trying a variation of a famous study conducted over 100 years ago by Sir Francis Galton. In 1906, he visited a livestock fair, and noticed a contest in which an ox was on display. Visitors were invited to guess the animal's weight after it had been slaughtered and dressed. Around 800 people participated, but not one person got the right answer of 1,198 pounds. However, in his article Vox Populi (http://wisdomofcrowds.blogspot.co.uk/2009/12/vox-populi-sir-francis-galton.html), Galton noted that "the middlemost estimate expresses the vox populi, every other estimate being condemned as too low or too high by a majority of the voters". In modern terminology, this is known as the median and in this case was 1,207 pounds. What was surprising was that this was within 0.8% of the weight measured by the judges. The mean of the guesses (1,197 pounds) was even more accurate! This phenomenon has become known as the Wisdom of the Crowd.

However, subsequent researchers have cast doubt on how generalisable this effect is and whether crowds can truly be wise. James Surowiecki has published a very readable summary of the research, called (unsurprisingly) 'The Wisdom of Crowds'. In it, he suggests that one of the requirements to ensure "crowd wisdom" is independence of the guesses. In other words, no conferring between members of the crowd so that social influences can cause biases in the estimates. What we wanted to explore by holding this competition was whether other guesses would have an effect on an individual's guesses and by how much.

These are some of the things we found so far:

- We had 219 contestants in the competition. 

- Of those 219, 67 (nearly 31%) people changed their guesses as a result of seeing the other guesses.

- The mean of all your initial guesses was 1145 grams. This was 15% more than the actual weight.

- The mean of all the changed guesses was 1025 grams. This was only 3% more than the actual weight, so much closer to the truth. This is contrary to Surowiecki's suggestion in that the influence of other's guesses actually had a positive influence on the overall result. We believe that this is because if an individual guesser has made a very extreme guess, they will then re-evaluate their estimate in light of information from other guessers to be something closer to the previous guesses.

What's next? Well, we also have kept the order of all the guesses so that we can attempt to see how the estimates change over time as well as from the influence of the previous guesses. We will update this page in due course with any interesting findings.