Why I give my students a ‘tragedy of the commons’ extra credit challenge

The professor omitted the sole interesting number: what percentage of students choose 6 points. Is it typically 15% or is it typically 80%?

It makes a huge difference. Most likely, it's a percentage low enough to indicate that the 10% threshold is carefully chosen. Otherwise he'd tell us.

This is evidence that he has an agenda to promote (that people are too selfish), and that the facts don't support his agenda (most people actually cooperate). So he omits the facts.

Actually, I think in this case the only rational choice is to ask for the 6 points. One of two things will happen:

1) Less than 10% will follow suit, in which case you'll get a 6 point boost, and you'll have secured an advantage over 90% of your classmates.

2) More than 10% will follow suit, and noone will get anything, in which case you will have no advantage over your classmates, and neither will any classmate have an advantage over you.

If you select 2 points, then either:

1) Less than 10% will opt for 6, in which case you'll have no advantage over at least 90% of the class, but some members of class may have a 4 point advantage over you, or

2) More than 10% will opt for 6, in which case the outcome is identical to choosing 6 in the first place.

Because the "points" have no intrinsic value except in comparison to your classmates (unlike a true tragedy of the commons, where the scarce resource such as food, etc, has real value), there is no reason not to roll the dice.

EDIT: Many people think the fact that the course is "not graded on a curve" makes a difference. It does not, at least considering the class in isolation. I took courses in college where an A was pretty much a dead certainty, and other courses where getting a B was actually quite an achievement. In all cases my performance within a course had to be judged based on my performance against the other students within that same class.

It's funny how many people are saying "oh just get everyone organized and have a lottery/auction". Yes, if we could perfectly organize large groups of people we could solve the tragedy of the commons. Unfortunately, organizing large groups of people is hard.

Even in the scenario where everyone is a rational actor (ie, fantasyland), effectively communicating to a group of > 20 people and obtaining unanimous consent is an exceedingly difficult problem.

> I’ve been giving it to students since 2008, and only one class has successfully mastered the challenge.

That's a real surprise. The 2 points option is the safe one, so everyone should go with that. Sure, I could assume that all (or most) of the other students think so and just get the 6 points myself. But then again, chances are that the other students think the same. So let's just go with the safer 2 points.

Cooperative behavior is particularly difficult when the stakes are very high. In the context of climate change, increased energy costs have very real impacts on people day to day, particularly in developing countries. For this reason, it's difficult to imagine those countries would behave cooperatively (even if a figurehead signs a document promising they will). It may be that the better strategy is to have the best grades in the first instance, knowing that noone will get extra credit. That is, build a strong and resilient economy so that you are better prepared to deal with, and adapt to, the inevitable challenges associated with global climate change.

I think this would be more interesting if there was a way to explicitly declare the marks post-curve. In the event the distribution of marks is smoothed to any sort of typically college distribution it's entirely rational to just select 6 points and not care whether 10% of the others do or not. Selecting 2 leaves you open to a worse result on a curve, but selecting 6 is nearly risk free if the distribution of marks is precisely controlled anyway. 2 marks when the average pre-curve goes up more than 2 and the average needs to be fixed post-curve is a really crappy proposition.

I wonder if jealously could also play a role here. The option to choose 2 points doesn't seem completely risk-free to me from an "emotional" point of view:

Suppose you choose the 2 and the class was actually disciplined enough that the points weren't lost. Then you might think that there might very well be a small group of students who got 6 additional points because they were greedy and you were not.

If students underestimate the risk of all points getting lost, they might in turn overestimate the "risk" of getting less than others if they play it safe.

Get together, auction the 6 point slots. Then, depending on whether it works or not, either evenly distribute proceeds to everyone or return the money to the auction winners.

I wonder if the prof has given any thought to what made that one class unique, if there were any factors that distinguished it from his other classes

off the top of my head, I wonder if it was a particularly small class

though I could imagine a lot of other potentially relevant factors

This really doesn't work for me as an experiment.

10% is too easy to hit. Do 3 or more students need the 6 rather than the 2? Why should I choose the 2 if it won't do me any good?

In any class I teach, there are generally 10% on the line between grades that requires that I take into account a little more than just scores. Did they do all the homework? Did they improve on assignments or were they blowing them off? etc.

Given that 10% probably need the 6 and not the 2, this is likely to never work.

I had a tragedy of the commons that I hadn't intended. I had an extra credit programming assignment, but there was a catch--the students had to write the tests for the automated tester. If nobody wrote tests, nobody had the chance to get any extra credit. They were allowed to collaborate on the tests, but not the solution.

Everybody complained bitterly when nobody wrote tests.

My response was not terribly charitable: "This is extra credit, folks. Not charity. You all had two options.

First, write the tests yourself. Yes, writing tests is time consuming, but not horribly difficult. Some students started, realized the amount of work, and stopped. However, once somebody writes the tests, everybody get to benefit.

Second, encourage someone else to write the tests for you. Did any of you talk to the students most likely to write tests? Did any of you encourage these students to put in the work if you couldn't pull it off yourself? Did any of you offer to compensate them for their valuable time?

Many of you have been repeatedly asking for a group assignment since the beginning of the semester. THIS WAS A GROUP ASSIGNMENT AND YOU FAILED. It simply worked out that freeloaders weren't guaranteed to be carried by the best students. The best students realized the amount of work and simply opted out since they already had enough points.

Welcome to the real world. Figuring out how to get someone to do something that they don't really want to do is a fundamental problem of life."

But the challenge doesn't make sense, there isn't much risk, no fear of being known. Economic studies have covered this before, would you rather make $50k where everyone else makes $75k? or make $40k where everyone else makes $30k and people pick the latter because they want to have more than their peers. Everyone getting the same points doesn't change anything relative to those around me.

In a paper score, assuming 100, 2 to 6 points just doesn't mean much. 6 points is greater than 5, it could take me from 84 to 90! I'll go for the gusto and try for 6. 2 points just doesn't make any difference, and I wouldn't feel any better getting 2 points.

Sure, the course grades are not curved, but the students are definitely competing academically, especially those in the same class. There's class rank, honors status, merit scholarships, internships, job opportunities, etc., that are technically at stake. It makes sense to choose 6 every time.

I see it as a simple equation. I don't think it says more than that about people.

Can someone enlighten me as to why a 90% break point? Why not for example 50%? Is it because 90% introduces some level of collaboration or vice versa?

Sans collaboration or any other useful information (people I may already know in the class, etc.), I would say at best you have a 50/50 chance of a total stranger choosing the option that is better for the common good. That number is based on it being college age students. If it were the general public at large I would estimate at least 60% would choose the more personally beneficial option.

Am I just extremely pessimistic? Please give your thoughts.

Its funny because we see the tragedy of the commons "fail" (succeed?) all the time in national elections: people are so afraid to "selfishly" choose the candidate they actually want to win (and thus cause the opponent to win) that we all "cooperate" on a sort-of ok choice (just like cooperating on 2 vs 6). I guess that's argument for penalties.

I'm surprised no rich student ever just offered ahead of the tally that they'd pay everyone $x if no one choose 6, where x is more valuable than 4 points. He now has the ability to safely choose 6, and everyone else is paid for their cooperation.

Alternative experiment: every day, students are allowed to do this voting. Assuming that the first few days run like this experiment, I wonder if we'll see a change in their behavior. This actually models reality more closely, to some extent.

Perhaps it's just cultural differences, but nobody getting extra points sounds like the most desirable outcome to me. Perhaps in a setting where the exam doesn't count for anything it'd be okay, but it just seems so juvenile.

So you get all the grades from your fellow students and organise the bottom 9% to choose 6 and the rest of the class chooses 2

in 6 months 4 points on a random test will be completely forgot.. might as well game the professor that is trying to game you

The challenge was recently discussed on HN: https://news.ycombinator.com/item?id=9874215

This new article is from the lecturer's point of view.

I suspect that professors aren't upset that the game theory is validated, but rather they're upset that the students haven't done the analysis and thus choose to change the game by introducing structural changes.

"People conformed to the group’s positive behavior."

People are more likely to respond to social pressure, than to appeals to "do the right thing." Valuable lesson for startup marketing... and maybe even pricing.

Doesn't take into account the student's standing in the class. If I am confident with the quality of my term paper and believe it is worthy of an A, what do I care if I get additional points?

Really odd that the professor does not mention carbon taxes in the story. It doesn't make sense to borrow game theory from economics but completely ignore the idea of taxing externalities

Don't the students just get together and do a lottery? And in RL tragedy of commons, it's rarely anonymous.

what if you choose 6 and explain to your classmates that "It feels good to be cooperative both from a strategic and moral perspective" and the whole commons thing and convince them to choose 2?

The commons worked great in two situations--1) europeans and 2) villages.

Maybe if we could get back to small villages of european descent we could have a higher quality of life.