In 1907, Sir Francis Galton attended a country fair in Plymouth, England. The scientist and statistician was fascinated by a contest, in which a group of fairgoers examined a live ox and bought tickets to guess the weight of the ox's meat once it had been slaughtered and dressed by a butcher. Galton arranged to take the guess tickets at the end of the contest because he was curious whether the crowd (a mix of butchers and farmers as well as mildly interested townspeople) had guessed well or guessed poorly.
After analyzing 787 guess tickets, Galton discovered that the median guess, 1207 pounds, was just a hair above the true weight of 1198 pounds. Clearly, by pooling and averaging the efforts of each individual, a collective guess emerged that was highly accurate. This is a classic example of what we mean when we talk about "collective intelligence."
Honey bees also provide a wonderful example of collective intelligence. A single bee has a brain the size of a sesame seed, and isn't terribly intelligent. Just watching a forager trapped inside as she bangs into window glass over and over and over again demonstrates that. But working together, a honey bee colony is a very intelligent thing, and they manage to make very good decisions by averaging together the individual actions of each group member.
A major problem that a colony has to solve is where to send their foragers each day. In the multi-mile foraging range around a hive, there may be millions of flowers, in many different flower patches. How does a colony decide to send half of the foragers a mile to the east, and half of them a half mile to the southwest? They use their clever choreography!
Using the waggle dance language of honey bees the bees can send the right number of foragers to the right patches. Each dance that points to a patch is a vote for other foragers to visit that patch. The better the patch, the more bees will dance about it, and the more new foragers will be recruited. This sends the most bees to the best patches and a smaller number of bees to patches with fewer flowers or a lower concentration of sugar.
The waggle dance language of honey bees lets them communicate the direction and distance from the hive to a food source. A forager who dances 20 degrees to the right of straight up is telling her sisters to leave the hive and fly 20 degrees to the right of wherever the sun currently is on the horizon. How long each waggle run is indicates the distance that an audience member has to fly to find the flower patch.
An even clearer example of the collective intelligence of bees is their ability of swarms to choose the best possible nest site among various options. When a swarm settles on a tree branch, a few hundred nest site scouts will explore the landscape for potential nest locations. Over a few days, the scouts will dance to advertise the sites that they have discovered, while other scouts will follow the dances and investigate the possible nest sites themselves. The better a location is, the more dances it will lead to, and scouts will dance for the best site they've visited. This means that a swarm of only slightly intelligent bees can make an incredibly intelligent decision about where to build their new home, critically evaluating many different variables to choose the very best location.
Photo credit: Professor Tom Seeley
Dr. Seeley's work tested bee preferences for entrance size (pictured) as well as entrance orientation and position, box height above the ground, box interior dimensions, and other traits. Bee swarms make consistent decisions in favor of boxes with more desirable traits.
At Betterbee's annual customer appreciation Field Day event this summer, we held our own “Guess the Weight” contest. Guessers could lift a 10-frame medium super full of honey, as well as a full 10-frame deep super. The top three closest guesses for the combined weight of both supers won prizes (such as a 10-pack of assembled BetterComb) and our Director of Research and Education, Dr. Peck, snagged the tickets to try to replicate Galton's 1907 statistical study.
So what did we learn? The combined weight of the two supers was 127 pounds. Of 156 guesses, the mean guess was 115.3 pounds. The median guess was 107.5 pounds. The majority of guesses tended to be too low, and a few were far, far too high. The winner and runners-up for the contest actually guessed 128, 128.3, and 125 pounds - far closer to the true weight than either average. However, the average of the guess was still relatively close to the correct answer, showing that our "colony" of beekeepers and guests were still able to find a better average guess than the guesses that many individuals came up with.
So, our test wasn't quite as decisive as Galton's ox guessers. But this was still a good demonstration that a motivated crowd working towards the same goal can make a good collective decision (whether it's a crowd of beekeepers trying to find the weight of some supers or a "crowd" of bees looking for the best flowers or nest location).
Don't underestimate the ability of a large group (or bees or people!) to sometimes make a very good collective decision. And the next time you encounter a "guess how many jellybeans are in the jar" contest, grab your calculator and ask if you can examine the guesses before you submit your own.
Oh, and don't try to cheat at Betterbee's next Guess The Super Weight contest by entering multiple guesses with the same name and phone number written on them. You never know when a bored scientist is going to be carefully examining the tickets!
If you want to read even more about the notion of collective intelligence, we recommend Professor Tom Seeley's bestselling book Honeybee Democracy, as well as a book by financial writer James Surowiecki called The Wisdom of Crowds.
You can also read Galton's original scientific report of the ox guess-the-weight contest here.