The natural laws of human behaviour (not)
Unlike physical objects, which behave according to the laws of gravity or thermodynamics, our behaviour is not that easy to describe
--
If you have been accessing Wikipedia recently, you may have noticed a fundraising banner popping up. And has been the case for the last few years, the text seeks to encourage visitors to the site to donate using a technique known as social proof. Especially when we are uncertain how to act or behave, and we don’t have a strong preference, we will be inclined to look at what others do or don’t do, so the theory goes.
Taking the lead of our fellow humans
A frequently cited illustration of social proof comes from research by psychologists Noah Goldstein, Robert Cialdini and Vladas Griskevicius. They experimented with different messages intended to encourage hotel guests to reuse their towels, rather than require new ones every morning. Different formulations were tried, like a standard “Save the environment, please reuse your towels during your stay” or more direct “Save the environment, and join your fellow guests in reusing your towels — in a study conducted earlier, 75% of our guests used their towels more than once”. The participation rate for the former message was the lowest at 37.5%, while for the latter it was 44%. A variant personalized to the room ( “guests who stayed in room 313”) achieved nearly 50% participation — a third more than the standard message.
Charitable donations look like a prime candidate for the use of social proof. Many of us are a little ambivalent about giving money — yes, we want to be a good person, but no, we don’t want to give away our money just like that, for nothing in return. Being told that many others are donating might just persuade us to do likewise.
And yes, we can see that the Wikipedia campaign message is personalized, just like at the hotel — “To all our readers in <country>” — even the day on which the banner is presented is part of it. But wait: it doesn’t tell us how many people actually donate, but…