Encouraging team-work while discouraging free-riding

Question

I teach several courses for computer science undergrads, including algorithms, programming and game development. In all my courses there are weekly homework assignments. I encourage my students to do the assignments in teams of 2 or 3, both to practice team-work and to learn from each other's ideas. Most of them indeed submit in teams. I have learned from my students that, in many teams, there are "dummy members" - team members that just write their name on the page without spending any work. The other team members are OK witih it, since it does not cost them anything - they do the assignment anyway, so they do not mind helping a friend. Additionally, these "dummy members" sometimes help them out in other courses.

I thought of several ways to prevent this "free riding". One is to require students to present their homework in front of the class. The "dummy members" will probably fail this presentation. The problem is that, in a large class, it takes a lot of time and it is logistically complicated. It also puts some students into too much mental stress.

Now I have thought of a new idea. In every weekly assignment, there will be 7-8 different questions to choose from; each team of n students will have to choose and solve n+1 questions. So a lone student will have to solve 2 questions, a pair will have to solve 3 questions, etc.

On one hand, the number of questions per student decreases with n. Therefore, if all team members share the work equally, they have an incentive to form large teams.

On the other hand, the total number of questions increases with n. Therefore, a dummy member is no longer a "free" rider - he/she costs the team by adding to their work load, so they have no incentive to let him/her write the name without working.

What do you think of this scheme? What other ways are there to encourage teamwork while discouraging freeriding?

The number of registered students is about 40-60 per class, though only about 20-30 come to class - the rest prefer to learn from home.

CONCLUSION: I have just finished teaching a course in game development based on the n+1 scheme. To further discourage free-riding, I also gave a bonus for students presenting their homework in front of the class. I explicitly said that the bonus is given only to the presenter and not to his/her partners. My idea was that, if only one partner works, then this partner will present and get the bonus points. To my surprise, almost all presenting students explicitly asked me to share the bonus points equally with their partners! For me, it is a strong evidence that there was little free-riding. It is highly unlikely that a student who did all the work alone will agree to give away the hard-earned points to a free-rider.

Answer 1

I assume that your goal is not just to catch and fail the free riders but to see to it that they actually do the work that will result in learning.

However, I'll note that having a goal of "sharing the work equally" is unattainable except over time and/or averaged over many different interactions.

The most radical suggestion here is to flip the classroom so that people do the active work of learning under your observation. Rather then passively "consuming" information in a group and then working to solidify it outside your view, the opposite occurs. Now, you can watch how people are contributing and you can encourage a true exchange of ideas and sharing of skills.

Note that for large groups flipping is much harder, requiring space and possibly help. But working in groups is easier to monitor than individual work in a flipped classroom, no matter the scale. People help one another and so there is less burden on the instructor and any TAs

But another idea that is less radical is to have students simply do peer evaluations of all group experiences. Students are reluctant to bad-mouth their friends, of course, so you need to use a scheme that they will consider "safe" in their social interactions. In pairs, you can ask two questions. First, what was your partner's chief contribution. Second, what was your own chief contribution. These are done separately. Note that both questions are positive and are less subject to simple opinion.

In larger groups, say five, I would ask to name the three most important members of the group and to also say why for each of them. The question about your own contribution is also needed. Note that a student who is never given praise is probably contributing less. But I didn't try to reduce these comments to numeric scores.

Knowing that others will have something to say about your contributions encourages them, I think.

To make this work you need to have a hand in forming groups so that the same people don't always work together. Different tasks can use different ways of forming groups, but some of them need to be under your control, even if it is just random assignment.

I would have students answer these questions on index cards so they can be sorted. Now you can look at a set of comments about each student and make judgements as needed, but you also get an idea about who needs coaching.

My own experience with this is positive. In one case, as student who was not standing out in the classroom was said by his peers to be an outstanding contributor to their project. I wouldn't have noticed it but for these comments and thought he was a bit of a slacker. In this case the work was done outside my view and so would have missed the interactions.

---

Note that both of these suggestions are for group work. One of my chief goals for groups was that team members could help each other learn. Not everything a student learned needed to come from me.

Answer 2

I’ve never heard of such a scheme, but it’s a cool idea and potentially worth trying.

One drawback I can think of is that it puts weak students at a disadvantage^* that could be seen as unfair since it compounds the disadvantage they already suffer from by being weak - a kind of “the poor get poorer” effect. If we imagine a weak student who isn’t lazy or a free-rider but is simply not very good at solving problems and thus isn’t expected to make a large contribution to the group’s problem solving efforts, we can expect that stronger students would not want to include such a student in their group since by including the weak student they increase their expected workload. (That is, although your scheme is designed to make large groups more “profitable”, in this particular situation the effect will be reversed). The weak student may therefore end up being forced to be in a singleton group by themselves, and therefore having to solve 2 homework problems per week when stronger students only have to do 4/3, 5/4 problems etc since they share the effort with other (also strong) students in their group.

That’s the general idea - this could make for a fun game theory model to try to analyze. But probably we won’t know how it works in real life until you or someone else tries it.