Collective Action Game: Failure in Action

So last night I ran a game that utterly failed.

It was in the first meeting of my upper level seminar on Environmental and Energy Security, which meets 1x week for 8 weeks in a 4 hour night class.  One of the key concepts I wanted to review was the Collective Action Problem, and naturally, my thoughts immediately went to how I could turn this into a game.

The game itself is below, but I thought it might be interesting to discuss how I created it and how it went so wrong.  Designing a game is a constant work in progress, and we shouldn’t be afraid to test it out on our students and see what happens.  It helps improve the game for the next time around, and I’ve always found that even when the game doesn’t go exactly as planned (or, you know, dies horribly while 20 undergraduates stare at you, expectantly), there is usually a relevant lesson for the players.  Luckily, that’s what happened this time.

Collection Action, in Action

The core of the collective action problem is trying to understand the ways in which individual self interest can preclude group interests.  If contributing to the group interest poses a cost, individuals who may value that collective goal may still fail to contribute, perhaps hoping to free ride on the efforts of others.  This is particularly a challenge in larger groups, when failures to contribute may be less likely to run social risks than in smaller groups, where contributors can be socially rewarded and shirkers, socially punished.

What I wanted, then, was some kind of decision-making game where students had to choose between a self-interested option and a collective option.  I quickly zeroed in on extra credit as the self-interested option.  The challenge here was figuring out how much extra credit was enough to tempt the students away from the collective option, without being too high.  I went with 20 points– my course is out of 10,000 points, so this seemed reasonable at the time.  As we will see, it was far, far too low to incentivize the students to behave selfishly.

For the collective interest, I considered two options.  One was some kind of group prize, such as baked goods. The other was some higher amount of extra credit.  What I did decide was not to tell the students exactly what their collective interest would be.  This was because my class was full of former students of mine, many of whom have tried to break my games in the past. Limited information was therefore my friend, or so I thought.

For game play, I also wanted to see what role social norms would play.  That meant playing the game anonymously the first time, and then openly the second, to see what kind of impact it would have if they would have to ‘out’ themselves as self-interested players.  This part also did not turn out as planned, by the way.

Here’s what I did:

The Collective Action Game

In round 1, students had to take out a piece of paper, write their names on it, and then indicate whether they chose option A, which was an automatic 20 points of extra credit on their final course grade, or option B, which was a contribution toward the unknown group prize.  75% of the students had to pick option B for the collective good to be provided. Choices were kept completely anonymous, but I announced how many students chose option A and B.

The collective prize was 100 points of extra credit–reduced by 20 points for each student who chose option A (basically, they were taking those points out of the collective pool).  In my class, 4 out of 20 students chose Option A, so that left 20 points for the collective prize.

Round 2 was exactly the same, except now Option A was worth 80 points, and the unknown collective prize was baked goods for our next class. Again, I did not tell them the prize until after the results were announced.

Round 3 had an 80 point option A and a 200 point extra credit prize, but was done openly–self-interested players came to the front of the room, and collective players went to the back of the room.

So how did that go?

The Disaster

Well. Here’s what I expected to have happen: in the first round, most students would take the self-interested route, would be shamed by the collectively oriented players, and then in Round 2, the public round, most students would act in the collective interest.

That’s exactly what DIDN’T happen.  16 of my 20 students chose Option B in round 1.  Although they expressed anger at the 4 unknown students who chose Option A, many (though not all) agreed that they probably would have taken the selfish route if the incentive was higher. The public round 2 was supposed to provide contrast, but suddenly, there was no contrast to be had! So on the fly I made up a new Round 2, a private game with a higher extra credit incentive for the self-interested option.   Maybe 80 points was still not enough, though, as this time 17 students (1 MORE than before) chose Option B!

At this point, I was seriously concerned about the game failing entirely.  I’ve written before about game failure and how it can present learning opportunities, and we’ve discussed it on the blog, but that doesn’t mean it’s a a fun experience when you are facing a new class on the first night and your new, untested game is completely failing to work as planned. I was pretty much resigned to having to admit to the students that the game didn’t work, but decided to try the public round anyway.

In the public round, remember, I had anticipated 100% compliance with the collective option. Imagine my surprise when the social norms I expected to make everyone choose Option B resulted in the exact opposite–8 students came to the front of the room, going for the 80 points, and as soon as the ones in the back realized that this meant the collective good would not be provided, they tried to run up front so they too could benefit.  Instead of promoting cooperative behavior, the ability to see what their peers were doing resulted in higher self interested action!

Total. Game. Failure.

By that I mean that I had created the game trying to engender a particular outcome (a rookie mistake, by the way), and the students behaved precisely the opposite of how I had hoped and planned.  The game, in other words, failed to produce the behaviors I was hoping to discuss.

But I had two things going in my favor. First, and less importantly, I know that failure still provides opportunities to learn, and I was able to make some connections with what did happen that could still illustrate the lessons I wanted.  More importantly, my students had NO IDEA that the game hadn’t worked as planned. I didn’t let on, and therefore they treated it as if I, the all-knowing instructor (hah!), had anticipated their every move.  That meant that during the debrief, they were searching for ways to make the game relevant to the content–and they found them.   They were able to teach themselves and their classmates all kinds of lessons about collective action, showing that even if they hadn’t learned what I’d originally planned, they still learned quite a bit.

Salvage Mission

We had a wide ranging discussion about a number of points. First, the ability to observe and monitor fellow group members matters.  Social norms matter–but if the norm is to be self-interested, then there is no danger in behaving that way.  Likewise, observation allowed students to change their behavior based on what their peers did–which is the essence of free-riding behavior.  In this case, too many students free-rode, preventing the good from being provided.  Another point this showed is that collection action is easier to achieve in a small, highly connected group.  There were only 20 students in my class, most of whom are junior and senior majors that have taken many classes together.  Their cooperation is, therefore, precisely what I should have predicted.  We were able to discuss what the game would have looked like if played in class of 300 or so, and they agreed the outcome might have looked different.  We were also able to discuss how the incentive informed their decision–noting the crucial role that incentives play in motivating their own personal behavior.

The game clearly needs some tinkering and refining, but I don’t regret playing it.  It will give us a common reference point for the rest of the term, and any opportunity to get the students active and moving about during a four hour class is welcome.  Ultimately it teased out several crucial points about collective action that I wanted the students to understand, even if it wasn’t the points I originally intended.  And it also re-taught me the crucial lesson that failure is just a starting point for learning, for both me and my students. I may have failed, but I’ll still call that a win.