The Student's Dilemma

Introduction

It is important to recycle because our natural resources are limited and our landfills are full. However, when you are sitting at your desk and the trash can is by your side, while the recycle bin is a five minute walk away, do you make the extra effort to recycle? Your one can will not save the nation's bauxite resources, and it is not going to overburden the local landfills by itself. Rationally, you should do what is most convenient for yourself, which is to drop the can in the trash. But if everyone were to refrain from recycling, our future could be in serious jeopardy (Huberman).

While such dilemmas involve many variables, at the root of the decision is the rational thought process which calculates the payoff for the individual involved. My project examines the factors which influence a person to act in the interest of the common good rather than for individual gain.

Analysis of the social dilemmas is based in game theory. The prisoner's dilemma is a classic example of the application of game theory, and introduces the buzz words "cooperation" and "defection." The scenario runs as follows: You and another person are arrested for committing a crime. After being isolated from each other, you are both offered a deal. If you maintain your innocence, and the other person does the same (you both cooperate), there is only enough evidence to put you away for two years. If you confess that the other person committed the crime (you defect) and he denies the charges (he cooperates), then he will serve five years in jail while you go free. The same results apply from his standpoint However, if you both accuse each other (you both defect), then you both spend four years in jail.

The prisoner's dilemma is a non-zero sum, two-person game, meaning that the interests of the two people involved are not completely opposed to each other; one person's gain is not directly the other person's loss. For games such as this, there is a calculated equilibrium point of the two players' strategies called the Nash equilibrium after recent Nobel laureate in economics John F. Nash who was one of the first to work with it. The Nash equilibrium is an extension of John von Neumann's minimax theorem which applies to zero-sum games.

For the prisoner's dilemma, the Nash equilibrium point is for both players to defect because neither person can justify cooperation. If you were to defect and the other person cooperates, then you go free, the best possible case. Should the other person defect, then it is a good thing that you defected, or else you would have ended up with the worse possible case, spending five years in jail. Of course the best case for each person would be mutual cooperation, but after you have rationally determined that defection is the best strategy, would you risk cooperating, knowing that the other person has done the same analysis? In this example, the Nash equilibrium has shown that two people involved in such situations are doomed never to realize the best case scenario as long as player strategies are guided by rationality.

The social dilemmas are basically prisoner's dilemma situations involving many people in which the majority of those involved are drawn to defection for the same reasons.

The Student's Dilemma

I call my social dilemma simulation "The Student's Dilemma." To begin, a teacher makes the following announcement to the class,

"I would like to offer all of you some extra credit. Write down on a piece of paper whether you would like to receive three or seven points. Everyone will receive what they ask for, as long as not more than 40% of the class asks for seven points - in which case, no one receives any points"

The individual student can only receive points if the percentage of students requesting seven points is under 40%. In a class of twenty-five students, one student's request is almost inconsequential. If the class is over 40%, then it doesn't matter what the student requests because he won't receive any points. Should the class be under 40%, though, the choice is between three or seven points. Rationally, the student would rather have seven than three point - but when everyone in the class makes this decision, they become locked in a situation where no one receives any points.

Putting aside the specific rules of the simulation, it can be agreed that in the essence of the dilemma, it would be better for everyone if each person cooperated (game theory lingo for working for the good of the whole). As cooperation is seldom the outcome, however, I wanted to find ways to reverse the trend. I hypothesized that if students knew they would be involved in such a situation many times, then they would be willing to cooperate, realizing they could improve their situation. Removing the anonymity of the dilemma, making people responsible for their actions, could motivate people to cooperate through guilt. Or, since people realize that cooperation is necessary for success, if they could talk with each other and have faith in mutual cooperation, then they would work for the good of the whole.

Eight different classes participated in "The Student's Dilemma": four Government classes (seniors; Groups A, B, C, D), two A.P. Biology classes Juniors and seniors; Groups E, F), and two Earth-Space Science classes (freshmen; Groups G, H). The simulation ran for one week (two weeks for Groups E and F). Group A was the control group in which students were not allowed to talk with each other, to know the responses of their classmates, or to know if the simulation would continue. The only information each student was given was the original instruction from the teacher, and the percentage of students who had requested seven points the previous day

Here are the exact instructions given to the teacher for Group A:

In preparation for the simulation, ask students to take out a sheet of binder paper, and at the top write the letter of their assigned group (A). Then assign numbers I through X for all students who turned in a permission slip^[1]. Have students record their number at the top of their paper.

Please make a list of student names and numbers (or put student numbers on their permission slips and I will make a list for you). Each day, I will collect the logs, and return them to you the following day with the percentage of students who requested seven points. Before running the simulation, please announce the percentage from the previous day.

(If a student doesn't understand the following directions, please reread the necessary passage, and try to use your own words as little as possible. This will help me in making sure that all classes are exposed to the same information.)

To begin the simulation, please read the following:

"In coordination with the science fair project which you turned in a permission slip to participate in, I would like to offer some extra credit to the class. Please do not talk or communicate in any way with the other students in the class about the simulation while it is in progress. Also, do not talk with classmates outside of class, or with other students who are not in your class about the simulation. Other classes my be involved in similar simulations, but the rules for each class are different, so you must listen closely to the rules specified for your class."

On your log sheet, write down whether you would like to receive either three or seven points; no intermediate amount requests will be accepted. Everyone will receive the amount they ask for, as long as not more than 40% of the class asks for seven points - in which case no one receives any points. Next to your request on your log sheet, explain why you asked for the amount you did. (After the first day read the following, "Write the date and a new entry for each new day.") If you write "I need the points" for your explanation, explain why your request will help you more than the alternative. Remember your responses are confidential. Keep in mind the fact that you are making your decision independent of those around you, but there are many others making the same decision.

The next seven groups followed similar rules, except that different variables were introduced for each. In Group B, students were told that the simulation would continue "after today" (iteration). For Group C, the names of the people who requested seven points were made known to the class (responsibility/guilt). Students in Group D were allowed to talk with each other during the simulation (communication). Group E was a combination of the iteration and the responsibility/guilt variables, and group F was a modified iteration where students were told "the simulation will last for two weeks, ending Friday." Group G included the iteration and responsibility/guilt variables, and Group H had iteration and communication.

The numbers used in the simulation were not chosen through any formula, but the amounts of three and seven points had to be important enough to the students for them to take the simulation seriously, successfully placing them in a social dilemma situation. The 40% was used so that it would be hard for the class to obtain, but just high enough for the percentage of defectors to be annoying to the cooperators.

Conclusion

Effectiveness of Variables

As far as inducing cooperation among students, the communication variable was the most successful. In Group D, the students didn't begin by talking to each other, but after the second day when the percentage of defectors reached 50%, one student began organizing the class's point requests. Each day it would be agreed that a certain number of people would be allowed to request seven points, and before turning in the requests, the class would calculate the percentage of students asking for seven points to make sure they were below 40%. Then, in perhaps the best solution for this specialized situation, the students would rotate who would be allowed to request seven points.

Group H started out by cooperating through communication. Again it was one person who rallied the class together, and convinced them all to request seven points. But the results the next day, revealed that 22% of the class had requested seven points. Some students had realized that with the entire class agreeing to ask for three points, they could get away with requesting seven points. Seeing that people they had trusted had defected, the tendency of students to cooperate declined over time. By the end of the week, the student's weren't talking to each other, and the percentage of student's requesting seven points was well above 40%.

In Groups C and E where the responsibility/guilt variable was used, the majority of the students didn't mind having their names read to the class, presumably because of their rugged upper-class egos. For the freshmen though, Group G, the percentage of students gradually decreased over time. These results indicate that for the freshmen students who were perhaps more concerned about having their names read to the class, the responsibility/guilt variable was effective in dealing with the individual's dilemma.

Another interesting trend in the groups where the responsibility/guilt variable was used was the difference between the percentage of male and female students requesting seven points. In Groups E and G, the percentage of females requesting seven points was about 15-20% less than that of male students. As in the case with the freshmen, this data shows that female students cared more about the class's opinion of them. This same gap between male and female requests also occurred in the control group, though, so perhaps further experimentation would show that in general females are more interested in working for the good of the whole or feel more guilty about hurting the class than males..

Questionnaire

The following questionnaire was given to students after the last day of the simulation:

Please answer the following questions. (Remember all responses are confidential.)

Gender?: M F

Grade in this class before the simulation?:

1. There is a water shortage, and you are asked by the city, but not legally bound, to only bathe on two certain days a week. Today you have been working on the plumbing on your house, and you really need a shower, but you have already used your two days to bathe this week.^[2]

Do you:

a) shower

> feel good, take a little bit of water from the city

b) don't shower

>feel icky, spare the city a little bit of water (a) or (b)

Why?

2. You go out to lunch with your friends, and you agree to split the bill so as not to over burden the waiter with multiple bills. (Your friends cannot tell by looking at your plate the price of the meal; after lunch. all each of you knows is the portion of the bill that you must pay.)^[3]

Do you:

a) order a higher priced meal

> you eat a higher priced meal and pay slightly more overall

>you eat a normally priced meal and pay a bit lower overall bill (a) or (b) ?

Why?

3. You are sick with a cold and are still contagious enough to give your cold to others. The last time you stayed home, you fell behind in school and lost money from hours lost at your job. Staying home today will have the same results.^[4]

Do you:

a) go to school and work

> keep up in your classes, make money, increase chance of getting others sick

b) stay home

>get behind in your classes, lose money, don't risk getting others sick (a) or (b) ?

Why?

4. A national charity that you believe in is asking for donations.

Do you:

a) donate

>your money is virtually nonexistent in the mass of funds collected by the charity, you are personally out a bit of money

b) do not donate

>the charity does not notice that you did not donate, you are not out any money (a) or (b) ?

Why?

5. You are a frequent buyer of a certain product. The product is superior and less expensive than the nearest competing product. You discover that the company that you buy from does not provide its workers in third world countries with safe and healthy working conditions. You do not agree with this.^[5]

Do you:

a) continue to buy the product

> you are not inconvenienced, you support of the company is virtually unnoticed by the company

b) don't buy the product

>you are inconvenienced, you lack of support for the company is virtually unnoticed (a) or (b) ?

Why?

6. You and another person are accused of a crime. You are offered a deal, which you are told is made to the other person, but you must come to a decision without communicating with the other person. The deal is as follows:

a) you accuse the other guy

_ if he denies the crime, you go free and he gets 5 years in jail

_ if he accuses you, you both get 4 years in jail

b) you deny the crime

_ if he denies the crime, you both get 2 years in jail

_if he accuses you, he goes free and you get 5 years in jail (a) or (b) ?

Why?

Did you violate any of the rules of the simulation? This may have included communicating with other students in class (if it was not allowed) or out of class, or looking at the responses of other students (if not allowed). Please explain how your violation of the rules influenced your requests in the simulation.

Many of you said that you chose 3 points because you didn't need the points. I you did need the points, how would your requests have been different, if at all?

Have you been influenced in this simulation by prior exposure to this subject (through myself, books, magazines, etc.)?

Did you participate in a simulation in another class? If so, what class, what period, and under what student number?

For the first question, according to the students, the value of hygiene outweighs the little use of water that is an exception to normal conservation; 73% of students chose (a). In the second question, people chose to order a regularly priced meal because it was one of their friends who would be stuck with the higher bill. Some people who chose (b) said that if they were with people other than their friends, they would chose (a). Of the 17% who chose (a), many people explained that they wanted the best meal they could get regardless. There was no real consensus for the third question. People who chose (a) said that their school and work were more important that the possibility of getting others sick. Those who chose (b) were mainly interested in their own recovery; a few didn't want to get others sick.; 64% of students chose (a). The large majority of students felt that because it was a cause that they believed in, and "every bit would help" they would donate. Those that didn't want to donate said that they didn't have money to give. Perhaps these results would have been different outside of upper-middle class Los Gatos; 71 % of students chose (a). For question five, those who chose (a) said they didn't care about the workers and those who chose (b) said they did care; some said they would try to influence others to broaden their impact; 56% of students chose (a). For question six, many said that whether they actually committed the crime or not was an important point. Presumably this was because they would take their punishment if guilty or go free if not - but what if the payoff was real? Of responding students, 37% chose (a).

Social Dilemma Behavior

One of the most important results of the simulation was that spontaneous cooperation did not break out in any of the classes without the help of variables. This clearly shows that when people are in social dilemma situations, they will most likely work for their own individual gain rather than the good of the whole.

In deciding whether to ask for seven or three points, the value of the points to the student was a very important factor. In student logs, the most common explanation for requesting seven points was "I need the points," and about half the time when students requested three points, they explained it was because they wanted to give others the opportunity to receive seven points. As the payoff was to improve grades, it is fitting that the AR Biology classes, Groups E and F, where the students were more academically competitive, had the hardest time working for the good of the whole. The value of payoffs is also reflected in the questionnaire results where students cared more about good hygiene than eating an expensive meal.

There were three recognizable types of students in the simulation: those who were unchanging in their requests, those who were constantly changing their requests, and those who made requests based on a set strategy. Students who were unchanging in their requests comprised about 5% of all students in all classes. It was more common for students to cooperate continuously, but there were those who asked for seven points each time. One student explained that he wanted to try to keep the class from getting extra credit by requesting seven points each day. The great majority of students were swayed in their requests by the daily reports of the percentage of students requesting seven points, trying to guess if there would be room for them to request seven points. There were also those students who realized after a few unsuccessful days that they must change their requests from seven to three points in order to help bring the class below 40%. Likewise there were also students who felt that they had "done their part" after the first few days and decided it was time to get their share of the points by starting to request seven instead of three points. A few students had strategies such as alternating requests for three and seven points, or alternating requests for seven points with a partner (in a class where communication was allowed.)

Limitations

There were limitations in the simulation. Over a time period of one week (two weeks for Groups E and F), it is hard to monitor a real development of trends. Perhaps over a longer period of time, the iteration variable would have been more effective. Of course for my experiment, running the simulation for long periods of time could offset the grading scale in the class more than would be acceptable for the teacher, which was why I was limited to one and two weeks

The fact that each class was made up of thirty or less students also affected the simulation because it made each student worth about 3-5% where in a real life social dilemma situation, each person is an imperceptible percentage of the whole. This may have convinced people to work for the good of the whole more often because their personal request could make a visible difference to the class. Breaking down large masses into smaller groups could be another solution to the social dilemmas. Few students were seen calculating their individual weight in the class, however.

When the rules of the simulation were announced to the class, the wording for the catch phrase was, "...as long as not more than 40% of the class asks for seven points." Putting the social dilemma in terms of defection may have made students think of asking for seven points as something that they must try to keep from doing. If it had been worded "at least 60%" then the simulation would have been more of a challenge to achieve, and perhaps it would have resulted in more people requesting three points.

While probably the least measurable of the many variables in the simulation, the socioeconomic make-up of the students should be noted. Los Gatos High School, where the simulation was run, is comprised mainly of upper-middle class Caucasian students. People who have had good things happen to them in the past are more likely to do good in the future, and perhaps living in the comfortable environment of Los Gatos made students more flexible to working for the good of the whole rather than for individual selfish gain.

Real-World Examples

The social dilemma situations are interesting because they can model many real world problems. Take the problem with America On-line (AOL) for example. With the same monthly rate to pay, whether the user is logged on for one hour a day or twenty-four hours a day, the individual customer is placed in a social dilemma situation. It would be more convenient for the individual and negligibly harmful to the community of AOL users to stay hooked up day and night. Only hooking up when the individual wants to use the Web would be more courteous to the rest of the AOL customers, but a lot less convenient for the individual since it may not be possible to hook up for some time due to space being used by other customers. Of course among the millions of AOL customers, the individual is virtually inconsequential, but when the rational decision to act for individual selfish gain is made, millions of AOL customers make it extremely hard for anyone to use the service (Mednick).

In the case of the AOL, the failure of cooperation results in a deadlock between participating parties. Most social dilemma consequences are much more serious, often ensuring the demise of the community. The tragedy of the commons is a well known example, first introduced by Garret Hardin in Scientific American in 1966. The tragedy of the commons is basically the social dilemma situation where limited natural resources must be apportioned between many different people or groups. The resources could be anything from shared grazing land (as in Hardin's example from 17th Century England), to deep sea fisheries, to old growth forests, to fresh water, to unpolluted air and soil. The natural resource will replenish itself at a certain rate, and as long as the participating parties take only a certain percentage of the resources at a certain rate, the resources continue to be available for everyone. However, when greed increases individual consumption, the resource is greatly reduced, and eventually disappears. And once the resource is gone, it is often lost forever.

Perhaps communication and responsibility are important first steps. After my experiment, I found out about some familiar sounding solutions in the real world. Representatives from New England fishing companies are now sitting down to talk about an agreed rate of fishing, so as to regain a healthy population; however, they waited until fish populations were at I % of normal. The Sierra Club takes action against environmental offenders by printing their names in their newsletter and making their deeds known to the public.

At the very least, social dilemma situations need to be recognized. Steps then need to be taken to remove the paradox for the individual or to avoid the dilemma altogether.

Special thanks to the participating teachers and their classes who helped make this project possible: Mrs. Joanne Benjamin, Mr. Les Kishler, Mr. Steve Hammack, and Mr. Steve Barth. Also, Mr. Dan Burns, Mr. Marc Mednick, Mr. Leon Felkins, and Mr. Bernardo Huberman for their input on the project.

Appendix

Notes

1. The Science Fair requires informed consent forms from all human subjects involved in science fair projects.

2.Felkins, "Cooperative Society"

3.Huberman

4.Mednick

5.Felkins, "Voter's Paradox"