Why grouping goes wrong
Most courses form teams one of two ways: students pick their own, or staff sort them by hand. Self-selection rewards the already-connected and leaves newer or quieter students scrambling. Hand-sorting is slow and usually falls back on whatever data is easy to reach, like who is in which tutorial.
Both work most of the time. Both also leave a few people in groups that were never going to gel, and you rarely find out until the project is well underway.
Start from what a fair team needs
A team that works tends to have a mix of strengths, some overlapping availability, and at least a bit of say from the people in it. None of that is visible from a class list, so the only way to form on it is to ask for it first.
Balance, do not just sort
Fairness is not stacking all the strong students together, and it is not splitting friends up on principle. It is spreading skills and experience so no team is set up to fail, while respecting a reasonable number of the preferences students give you.
Give students a voice without handing over control
Asking who someone wants to work with matters, but mutual preferences should be one input among several, not the whole decision. The aim is teams people can succeed in, not only teams people asked for.
Review before you publish
Whatever forms the teams, a person should look before students do. The value of a draft is that you can fix the one team that looks off while it is still just a draft.
Doing all of this by hand, balancing skills and schedules and preferences across a whole cohort, is genuinely tedious. It is also exactly what Teamer automates: you set the inputs, it generates balanced draft teams, and you review before anything is published.
If you would like the survey question set I use as a starting point, get in touch and I will send it over.
Keep reading
The handful of questions that improve matching, and the ones that just add noise.
Group work and the students who get left behindGroup projects reward the well-connected by default. Here is how to change that without policing every team.
How many students should be in a group project team?There is no magic number, but there is a range that works and a few reasons teams drift outside it.
Should students pick their own teams, or should you assign them?Self-selection feels fair and saves you work. Here is what it quietly costs, and when assigning teams is the better call.
How to assign students to groups in a large classSorting a class list by hand does not scale. Here is a process that does, without turning into a spreadsheet nightmare.
How to run a team health check on student projectsMost struggling teams are quiet until it is too late. A lightweight check-in surfaces them while you can still help.
How to stop free-riders in group projectsFree-riding is usually a design problem, not a discipline problem. Here is how to reduce it before it starts.