Game Theory- Prisoner’s Dilemma in Real Life
http://www.businessinsider.com/prisoners-dilemma-in-real-life-2013-7
The Prisoner’s Dilemma might be one of the best-known models of Game Theory, including references to behavioral, psychological, and biological applications. In class, we discussed the predictions of the Prisoner’s Dilemma, coming to the conclusion that even though the best outcome for both prisoners would be to Not Confess, the more likely possibility is that both players end up Confessing, due to the fact that Confessing is a dominant strategy for both Prisoner’s. This “game” highlights the psychological aspect of game theory, and explores the human reasoning behind choosing to confess against another player.
According to the article in Business Insider, the Prisoner’s Dilemma has not been officially run as an experiment on real prisoners before. Two economists, Menusch Khadjavi and Andreas Lange, at the University of Hamburg decided to change that and run the experiments on both students of the University but also on Inmates of the Lower Saxony’s primary women’s prison. They ran two versions of the game, a simultaneous version (the version we learned in class) where you cannot react to a different player’s decision, but only to a guess on what their decision might be, and a sequential simulation, where a player would know the other players move and would get an even higher payoff if the second player decided to betray the first player if the first player choose to not confess. Instead of lengthening/shortening prison sentences the participants were given payoffs in money if they were students, and the equivalent in coffee/cigarettes for the inmates.
Khadjavi and Lange were expecting the results to be what we discovered in class, which is that the majority of prisoners would be more “distrustful” of their counterparts and therefore confess, and students would have less of this type of mentality so they might choose to not confess and end up collaborating with the other player more than the inmates would. They discovered that in a simultaneous game, inmates didn’t confess 56% of the time while students didn’t confess 37% of the time, allowing for 13% of student pairs and 30% of prisoner pairs to get the best mutual outcome. In the sequential game, students’ that didn’t confess went up to 63% so the student pairs ending up with the best mutual outcome went up to 39% while the prisoners’ numbers stayed about the same. This confirms what behavioral psychologists have been thinking for years, which is that a lot more goes into real-life game theory than simply the best calculated strategy. We can see that not only does a best response mentality have to be taken into account for, but in addition we need to factor in human emotions, and human responses to what they think others might choose. There’s a lot more than just calculated self-interest which has to be taken into account, feelings like empathy, and compassion have a lot of strength behind them as well.