What would happen if a prisoner’s
dilemma experiment were performed in real-life? Well, a Business Insider article presents the outcomes of an experiment that compared the cooperation of
prisoners in Lower Saxony’s primary women’s prison with a group of female
college students.
In lecture, Professor Shirk
demonstrated prisoner’s dilemma for our class with the payoff of candy and the
punishment of no candy.
In this experiment, rather than
offering years off of a prison sentence (or candy, like in GVPT200), the group
of female students received their payoff in Euros and the prisoners received
the equivalent in cigarettes or coffee. When pairs were given the option to
defect or cooperate, it was found that the prisoners were significantly more
cooperative. The article notes that 37% of students and 56% of inmates
cooperated. And 13% of student pairs and 30% of prisoner pairs cooperated with
each other.
The Business Insider reporter seems shocked by the results; however, I
would argue that the results are not so surprising. Prisoners typically live in
tightly-knit communities within their prisons. They often depend on each other
for cigarettes, luxury snacks/foods, and other rare commodities that the prison doesn’t
provide them. Also, the prisoners generally trust each other. Furthermore, this
experiment was more than just a lab game for the prisoners. When they return to the
prison, their behavior will have real consequences - they will either be punished or rewarded by their peers.
Since the prisoners depend on each other for the coffee and cigarettes they
receive in the game, they will probably be expected to share with their
friends.
In addition, some people in prison
spend their lives practicing game theory and playing prisoner’s dilemma. Also,
the criminal culture condemns non-cooperation. Many of the prisoners probably
have much more experience with how these games work than the average college
student. Their experience with and understanding of the game makes it likely
that they will be able to cooperate better than the students who have probably
never played a prisoner’s dilemma game.
I also think it is interesting how
all of the subjects, both prisoners and students, are all female. I think it
would be interesting to see how men would fair in this experiment, since men
and women typically behave differently in these games. Women stereotypically cooperate more than men.
One issue I find with this experiment
is that the payoffs and consequences are not as serious as risking years of
one’s life (as it would be in a true prisoner’s dilemma). To get more accurate
results, I would suggest the researchers observe the behavior of people who are
actually facing time in prison for a crime. They should also observe the
behavior of students who are facing more detrimental consequences, such as
academic probation for cheating. I feel that if the payoffs are more
significant, the experiment is more likely to yield results that closely mirror
the real world.
how do you think the game would fair if it was done with people who considered each other "brothers" such as fraternities or gangs? do you think cooperation would lower because of loyalty or fear of rejections? or do you hink deep down people would snitch about the same?
ReplyDeleteI assume that the female prisoners they tested were probably loyal to each other, much like fraternity brothers. I believe the college students they tested did not know each other, and they cooperated much less. This experiment from the article would argue that as loyalty increases, so does cooperation, since it gives both players a mutually beneficial payoff. However, if the members are very loyal to each other and face dire consequences for colluding (like gang members), they are probably much more likely to stay quiet and not snitch.
DeleteI agree that tested male prisoners may end in completely different results. Females consider emotions more when making decisions and may be more loyal than men. Females may be more willing to trust each other than men. I personally don't think they would be expected to share with their friends. Prison can be a very hostile environment and often times you have to fend for yourself. I think testing students would be a very interesting situation. Do you think there would be a difference in male and female students?
ReplyDeleteI definitely think there would be a difference in the way female and male students respond to the prisoner's dilemma. Not to be too stereotypical, but women are usually more cooperative than men. I think the we would see higher cooperation among female students and more snitching and looking out for self-interest among the men.
DeleteI understand that the people doing this experiment were trying to control for confounding variables by testing only females. But in doing so, they aren't really giving the full picture of how people react to the prisoner's dilemma if they're only testing women.
Very interesting Sara.
ReplyDeleteOne thought. Do you think that the prisoners are self-selected to be more cooperative? In other words, they may be the ones that either have been the sucker or are in jail because they decided not to rat. Also interesting is that students were less likely to cooperate than inmates. Maybe we should throw you all in jail!
I hadn't considered that before you mentioned it, Professor Shirk. But it's likely that the prisoners are self-selected to be more cooperative. I'd also assume that the prisoners have more experience than the college students with playing prisoner's dilemma-style games. Therefore, they probably have a better understanding of how to achieve mutually beneficial payoffs.
Delete