How to Split a Class into Groups
Splitting a class into groups is one of the most frequent tasks in teaching and one of the biggest time sinks. Let students sort themselves out and someone is always left over. Assign the groups yourself and you get accusations of favouritism. Random assignment reduces both problems, though it is not always the right answer.
Why grouping is genuinely hard
The difficulty is social, not mathematical. When students pick their own partners, they cluster around the popular ones and somebody is always chosen last. Those few seconds can set the tone for the rest of the lesson.
Assigning groups yourself creates a different problem. Even when the assignment has a good reason behind it, students cannot see that reason, so they read it as a judgement about them. Random assignment removes both failure modes at once: nobody is left unchosen, and the result visibly carries no intent.
The important part is not announcing that it was random but letting the class watch it happen. Putting the draw on the projector and pressing the button in front of everyone cuts complaints dramatically.
Group size: three, four, or five
Group size should follow the activity. Three makes free-riding nearly impossible and gives everyone airtime, but a two-against-one split isolates the odd person out. It suits practical work and short tasks more than discussion.
Four is the safest default. It splits cleanly into pairs and divides into four distinct roles. The trade-off is that three people can carry the work, so a disengaged student has somewhere to hide.
Five or more suits presentations and build tasks where roles are clearly separated. The risk is that a quiet student can go the whole lesson without speaking, so pair large groups with assigned roles.
- Three: practicals, short problem sets, when everyone must speak
- Four: discussion, general collaborative tasks, easy switching to pairs
- Five or six: presentation prep, build activities, clearly separated roles
Three ways to handle the leftovers
Class sizes rarely divide evenly. Twenty-seven students in groups of four gives six groups and three spare. You have three options.
First, spread the extras one per group, giving three groups of four and three of five. This keeps the group count fixed and holds the size difference to one person, which works well most of the time.
Second, form one extra smaller group. That adds to presentation time, but a group of three is denser and often works better for hands-on tasks. Third, give the leftover students a floating or recording role. If you do this, assign the role randomly so it does not land on the same student every week.
The team maker on this site takes the first approach automatically when the numbers do not divide evenly, keeping every group within one member of the others.
When purely random is the wrong call
Random is not always best. In a lab where safety or equipment handling matters, every group needs at least one confident student, and pure randomness does not guarantee that.
Long projects are the same. Over several weeks, an uneven spread of ability turns directly into an uneven spread of outcomes. Stratified assignment helps here: split students into tiers first, then randomise within each tier. With eight experienced and sixteen inexperienced students, you can guarantee one experienced member per group while the specific placement stays random.
Sometimes two particular students need to be kept apart. Running the draw and redrawing when the condition trips is far easier for a class to accept than a hand-built arrangement.
How often to reshuffle
There is no fixed answer, but there is a principle. If the activity finishes within one lesson, redraw every time. Students meet different combinations and group dynamics never calcify.
For a multi-week project, do not reshuffle mid-way. The cost of re-establishing roles and catching up on progress outweighs the benefit. Two or three reshuffles per term, aligned with project boundaries, works better.
Random redraws also mean the same pair will sometimes land together repeatedly. With twenty-four students in groups of four, two specific students share a group about thirteen percent of the time, so across eight lessons one repeat is expected. That is probability, not rigging.
A running order that saves time
A fixed sequence makes this smooth in practice. Confirm the list with absences removed, set the group count where the class can see it, then run the draw in front of them.
Once the result is up, assign each group a physical location before anyone moves. Saying go find your group without naming where costs several minutes. Tying group numbers to areas of the room makes the transition much faster.
Save the result as an image if the same groups continue next lesson. It saves the inevitable round of asking who was I with again.
Try this tool
Running it once makes the idea in this article click a lot faster.
FAQ
What should I do about absent students?
Remove them from the list before drawing. Leaving an empty slot makes that group smaller and changes how much work each member does. The team maker lets you edit the name list directly, so you can delete absentees and run it immediately.
I need to keep two students apart.
If the draw puts them together, run it again. Redrawing reads far more naturally to a class than editing the result by hand. With a larger class you can also place those two in different groups first and randomise everyone else around them.
Should I set group count or group size?
If the activity fixes the number of groups, such as available equipment or presentation slots, set the count. Otherwise set the size. The team maker distributes members evenly once you give it a team count.
What if students do not believe the result is random?
Put it on the projector and let a student press the button. Watching the process is what convinces people. The team maker shows the balls physically mixing, which suits this purpose particularly well.
The same students keep ending up together.
With twenty-four students in six groups, any given pair meets about thirteen percent of the time. Over a term, two or three repeats are entirely normal. If you want to avoid it, redraw only when that combination appears.
Are student names stored anywhere?
No. Everything is computed inside the browser and the names you type never reach a server, so there is no student data leaving your device.