Is cooperation natural?
This is the season for "job talks," a ritual at research-oriented universities where faculty members hear presentations from the best of the newly minted PhDs who have applied for jobs as new assistant professors, and then choose which to hire. The hopefuls discuss their dissertations during these meetings, so in today's blog and on Thursday, I will report on the research of two young scholars applying for jobs at Harvard Business School and at the Harvard Kennedy School.
Dave Rand -- whose PhD is in mathematical biology from Harvard, but who is nonetheless on the job market to work on game theory, negotiations, and such -- presented a line of research directed at the question: Is it natural for people to cooperate with others?
He did a series of lab experiments involving a kind of cooperation game that is often studied in the lab. The basic structure of the game is that if a person behaves in a cooperative way and the other person in the game cooperates also, then both players gain -- but if you act selfishly and the other person cooperates, you are even better off. That creates an incentive for each party to act selfishly, hoping the other will cooperate, but if both act selfishly, both suffer.
What Rand wanted to investigate was the relationship between how quickly people make a decision how to behave and whether they choose to cooperate or to act selfishly. His idea was that if fast decisions are more likely to be cooperative, then cooperation is natural, while if fast decisions are more likely to be selfish, then trying to take advantage of others is natural.
Since many people erroneously believe that most social science research simply establishes the obvious, I will say that before I heard the results of the experiments, I didn't have a prediction. Readers may wish to venture a prediction (and think about how confident you are about your prediction) before reading the next paragraph.
OK, so what Rand found was that people who decided faster were more likely to cooperate than those who decided more slowly. He then went on to review records from earlier research by other scholars using this game where decision speed was noted (though not used for research purposes), and found the same thing. He then tried a different method: He randomly divided up the group into three smaller groups, each with different instructions. One group was to decide in fewer than 10 seconds, another in exactly 10 seconds and the third in 10 or more seconds, with no upper limit. He found the same results. This seems like some handy information for somebody managing people -- or indeed for any of us -- to know.
So here's the next question: Will he get the job? The answer is TBD.
Posted on Jan 17, 2012 at 2:54 PM