Randomly assigned teams

From MathWiki

Discovery projects and group work

Georges Monette

About twenty years ago, an undergraduate student in a senior statistics course was applying for a job with RBC and he approached me for permission to give my name as a referee. He had very good marks and contributed a lot to the class so I said that I would be happy to be a referee.

A few days later I got a call from someone at RBC who wanted to ask me about the student. I started telling her about his good marks and she interrupted me. She wanted to know whether he worked well in a team. I answered that my course didn't include any team work. The interviewer replied that, in that case, I had no useful information for them. She thanked me and hung up.

This doesn't mean that RBC is not interested in academic performance but, from me as an instructor, they wanted to know about team work.

The next year I started experimenting with group work. I'm still looking for the ideal formula but I think that group work is a very effective component of courses – not just for its own sake but also because it promotes a kind of discovery that is specifically important in learning statistical concepts. In the early years I let students form their own groups but I found that they tended to stay in cliques of similar students. Also, I felt uncomfortable that some students seemed to be marginalized for various reasons and had trouble finding groups that would welcome them.

About ten years ago, I decided to form instructor-selected groups. Students complete a questionnaire at the beginning of the course and I use their responses to semi-randomly create diverse groups, spreading around, as much as possible, students with skills in computing or mathematics, students who seem adept at written expression, upper-year students who have more experience working in a university setting, mature students who can share their life experience with younger students, etc.

A very practical benefit of experience with team work is preparation for the work place and for job interviews. Almost all jobs require you to work effectively with teams you didn't select. Learning to work with people you don’t know is almost as important a skill as statistics and here's a chance to practice. To face the painful truth, being good at team work is actually more important than statistics.

When I run into old students, they often report that working with a group was one of the most rewarding experiences in the course and friendships formed in groups lasted long after the end of the course. Moreover, their experiences with group work and with interesting projects gave them a wealth of material to talk about in interviews, which they felt helped them a lot to land their first job.

Group work is particularly suitable for exploratory projects that require reflection and multiple modes of reasoning. The interaction among group members lets each individual discover approaches that they would not have thought of if they had worked alone.

This paper by Oakley, Felder, Brent and Elhajj (2004) "Turning Student Groups into Effective Teams (http://www4.ncsu.edu/unity/lockers/users/f/felder/public/Papers/Oakley-paper(JSCL).pdf) reports a study in which "students found by a two-to-one ratio that their worst group experiences were with self-formed groups and their best with instructor-formed groups". If anyone wants to follow up to find out more about the design of this study, I will very much enjoy finding out.