<Cross Posted with Ingenious Monkey>
Over Christmas I spent some time in Germany where I visited the Children's Science Museum in Cologne. The place was packed with kids who were enthusiastically participating in all kinds of experiments and hands-on research activities, so that I could not help but think to myself: "what a cool place this would be to collect developmental data. Think about it: Thousands of observation points for children of varying ages...and..."...
I guess I didn't finish the thought, but must have become distracted by some amazing perceptual illusion, or the free-throw shooting robot on the "computer and technology floor"...
Regardless of reasonable critique (click here for a German review of Cologne's Science Museum), I think that science museums are awesome, AND - so I just came to see in a recent journal article - they are indeed a nice place to collect developmental data from:
For a study in the journal Evolution and Human Behavior, researchers Peter Blake and David Rand from Harvard University set up a shop at the science museum in Boston and asked children ages 3 to 7 to participate in what economists call the dictator game; essentially an experiment in which children's sharing behavior could be observed. I always imagine that people who are unfamiliar with the dictator game must think that it is something rather diabolic (something like this), and so I really like the thought of these two researchers approaching children's parents and explaining that "we would like your child to participate in a few rounds of a dictator game...". Of course in reality the researchers certainly used a different approach line, but still, I like the thought...
Whatever the case, the dictator game is a classic experiment in game theory, in which one person (the dictator) receives an endowment and gets to do one of two things with this endowment: Keep all of it, or share some of it with a second person in the game (the receiver). Ice-cold rationality dictates that nobody should ever share their initial endowment, but warm and fuzzy experimental results constantly show that people share far more often than not; many times even sharing up to 50% of what they receive as an initial endowment. For example, in a study with American college students approximately 20% gave nothing, roughly 60% gave between zero and a half, and the remaining 20% gave half the stake.
Variations on the classic dictator game have shown the same behavior for different types of endowment (e.g. money or food), that it is possible to prime would-be dictators into be more altruistic, and that women tend to give more than men. With the particular variation that Blake and Rand carried out with the science museum kids, the goal was to investigate the development of prosocial behavior over different age brackets (i.e. at what ages do children start sharing in the dictator game?) as well as to investigate how prosocial behavior is connected to the value of the to-be-shared resource.
To do so, they provided participating children with a number of stickers; some of those were stickers that the children liked a lot, others were stickers that the kids were less fond of. Then they essentially gave children the opportunity to (anonymously) share any number of stickers with a (hypothetical) other child.
One of the findings from this study can be summarized as follows: The older the kids, the greater the likelihood that the children would actually share some of the stickers in the dictator game. For example, only 30% of three-year olds shared in the experiment, while more than 70% of all participating 6 year olds shared their stickers. Indeed the tendency to engage in prosocial behavior, so the study suggest, follows a fairly clear age trajectory along which children (age three to 7) become more prosocial as they grow older.
Ontop of this, the researchers found an interesting result regarding children's giving patterns in response to how highly they valued the stickers in the experiment. Here the data clearly shows that -regardless of age - for those children who decided to engage in prosocial behavior, the same giving pattern seemed to hold: Children who shared at all gave away about 40% of their favored stickers, and about 50% of the less preferred stickers. I.e. children were guided in their decision of how much to share by how much they valued the resource.
What makes the study really interesting (and the study design quite clever), is what you get from putting both of the above findings together: Since age brackets respond differently to the "task" of sharing, but respond equally to the task of how much to share when the endowment value changes, the study is able to provide compelling evidence that prosocial behavior comprises two distinct steps:
Deciding whether to give, and deciding how much to give.
According to the study's authors
"These two steps have different developmental trajectories, which implies that different processes are involved: the probability to give increased with age whereas children's level of donation did not change with age. The latter finding is notable because one might expect the fraction of stickers donated to increase with age, but this did not occur. Instead, for those who decided to be donors, the youngest children gave the same amount, on average, as the oldest children."
That posocial behavior can be broken down into two distinct steps, and that sharing might correspond to the value of the resource being shared, is of course not a new hypothsis - after all it follows from economic approaches to altruism and fairness -, but it is the design of analyzing children's behavior (in itself tricky) and the consideration of their development trajectory for prosocial behavior that creates empirical support for this idea in a way that has so far remained elusive for a number of similar studies with adult participants.
Related Post: Priming Toddlers to be Altruistic
Peter R. Blake; David G. Rand (2009). Currency value moderates equity preference among young children Evolution and Human Behavior : 10.1016/j.evolhumbehav.2009.06.012