When he encounters such a withdrawal of collaboration, the theory goes, the malefactor will learn the error of his ways and become a more co-operative individual. And there is experimental evidence, based on specially designed games, that tit-for-tat does work for pairs of people. Human societies, though, are more complex than mere dyads. And until recently, it has been difficult to model that complexity in the laboratory. But a paper published this week in the Proceedings of the National Academy of Sciences by Nicholas Christakis and his colleagues at Harvard has changed that. Dr Christakis arranged for a collaboration-testing game to be played over the web, with many participants. As a result, he and his team have gained a more sophisticated insight into the way co-operation develops.
Dr Christakis used what is known as a public-goods game for his experiment. At the beginning of such a game, points are doled out to each participant. During every round, players are given the opportunity to donate points to their neighbours. Points so donated are augmented by an equal number from the masters of the game. If everyone co-operates, then, everyone ends up richer. A “defector” who refuses to donate to his co-operating neighbours will, however, benefit at the expense of those neighbours. At the game’s end, the points are converted into real money, to ensure that proper incentives are in place.
To play his large-scale public-goods game, Dr Christakis recruited 785 volunteers via Mechanical Turk—a service provided by Amazon, an online retailer, that works by farming out small tasks to an army of individual workers. Each volunteer was randomly assigned links to, on average, eight other players. Together, they played repeated rounds of one of three variations of the game.
