Random groups that work: how to group students, with rules

When random groups help

Random groups are at their best when the point of the task is talking and thinking together, not producing a polished result. They help when:

• The class is new. Students meet people beyond their usual circle, and you learn who works well with whom.
• Friendship groups have set hard. A random shuffle breaks up the same four people at the same table without you singling anyone out.
• The task is short. A ten-minute discussion or a quick problem doesn't need carefully balanced teams.
• Fairness is in question. “Why is he always with them?” has no comeback when everyone watched the shuffle.

They're a poorer fit for long projects where each group needs a mix of skills, or where a particular pairing would genuinely be a problem. For those, choose the groups yourself, or use random groups with a few rules, which is where this gets interesting.

How big should groups be?

There's no single right size, but most teachers settle on a few rules of thumb:

• Pairs for quick “turn and talk” moments. Nobody can hide.
• Threes and fours for most group tasks. Enough ideas to disagree, small enough that everyone has a job.
• Five or more only when the task has clear roles for everyone. Otherwise someone usually drifts.

Rather than fixing a size, decide how many groups you want and let the sizes follow. WhirlPicker keeps group sizes within one of each other: a class of 27 in four groups gives 7, 7, 7 and 6.

Keep two students apart, quietly

Every teacher knows a pair who shouldn't share a table. The usual workaround, shuffling again and again until they land apart, is slow and visible. Students notice the re-rolls.

A better way is a rule. In WhirlPicker's group maker, add the two students as a keep apart pair. Every shuffle respects it, and the pair is remembered for that class, so you set it once. Nobody sees the rule on the board, only the groups.

Rules can conflict. If you ask for three groups and add enough keep-apart pairs, there may be no way to satisfy them all. WhirlPicker tells you which pair is the problem and offers a fix, like one more group or dropping that rule.

Leave out who's away

Groups made from the full register fall apart the moment you notice two names are absent. Mark who's away at the start of the lesson and they're left out of the groups that day, then included again tomorrow. It takes seconds and saves rebuilding a group mid-task.

Shuffle once, in front of them

Random groups only feel fair if the shuffle is real. Do it on the board, once, and stick with the result. If something truly can't work, fix it with a rule for next time rather than re-rolling until you like the answer.

It helps to know how many possible groupings there are. Split 24 students into six groups of four and there are 4,509,264,634,875 different ways to do it, more than four trillion. So when two best friends land together, it really is just chance. With odds like that, a “too convenient” group says nothing about the shuffle.

Give every group roles

A group works better when everyone knows their job. Four classic roles cover most tasks:

• Facilitator: keeps the group on task and makes sure everyone speaks.
• Recorder: writes down the group's ideas.
• Reporter: shares the group's answer with the class.
• Timekeeper: watches the clock and calls the next step.

Rotate the roles so the confident talker isn't always the reporter. The same idea works for whole-class jobs, which we cover in classroom jobs that actually rotate.

A quick routine

1. Mark who's away.
2. Choose the number of groups, not the size.
3. Check your keep-apart pairs are still right for this class.
4. Shuffle once on the board, then copy the groups into your slides or chat.
5. Give each group its roles, and go.