Random group generator

How to use the generator

1. Paste the names, one per line: a column from a spreadsheet, a register, or the participant list from a video call. Numbers and bullets are removed, and a heading line such as “Group A:” is skipped.
2. Choose how to split them: a group size (“Groups of 3”), a number of groups (“4 groups”) or pairs.
3. Decide about leftovers when you choose a group size: spread them over the groups, or keep one smaller group.
4. Share the groups. Copy them as text for a chat or an email, or as Markdown for your notes. “Shuffle again” makes a new set from the same list.

If someone arrives late, add their name: they go into a random place and everyone else stays in their group.

How the groups are made

The list is shuffled with the Fisher–Yates method, which gives every possible order the same chance, using the random numbers your browser keeps for cryptography. The shuffled list is then dealt into groups of the sizes you asked for. Here is what the settings do with real numbers:

Random groups or chosen ones?

Random groups are quick, and people see them as fair: nobody was picked last, and nobody can say you put them with their friends or kept them apart. They mix people who wouldn’t choose each other, which is what you want for discussion, warm-ups and short practice tasks.

For longer projects, choose the groups yourself. Self-chosen groups tend to form around friendships and leave the same people out each time, and a random draw can put all the experienced people in one team. Oakley, Felder, Brent and Elhajj (2004), in a much-used guide to student teams, recommend that instructors form project teams themselves for this reason. A good compromise: use random groups for anything that lasts a session, and chosen ones for anything that lasts weeks.

Group sizes for discussion and for projects

• Pairs for quick tasks and checks: everyone has to talk, and nobody can hide.
• Threes and fours for discussion: enough different views, small enough that everyone speaks.
• Three to five for projects. Larger teams need more coordination, and each person tends to put in less effort as the group grows, the effect known as social loafing (Latané, Williams and Harkins, 1979).

These sizes are rules of thumb from teaching practice, not fixed limits; the tool lets you pick any size. For breakout rooms in an online session, small rooms of three or four also make it less likely that a room falls silent.

Pair programming rotations

In pair programming one person types (the driver) while the other reviews and thinks ahead (the navigator), and they swap roles regularly; many teams swap every 15 to 30 minutes, though that is a habit rather than a rule. Rotating partners between sessions spreads knowledge around the group and stops strong pairs from pulling away from the rest.

Random pairs each session are the simplest rotation. Random pairs can repeat, though. If you need everyone to work with everyone exactly once, a round-robin rota does it: with 14 people that takes 13 sessions of pairs. For a few sessions, shuffling and swapping a repeated pair by hand is quicker.

Tips for running group work

• Show the groups on screen before anyone moves, and say where each group sits or which room it uses.
• Decide what happens to leftovers before you shuffle, so nobody becomes the odd one out in front of everyone.
• If two people really can’t work together, swap them by hand and keep it quiet. Reshuffling until the groups look right undoes the fairness that makes random groups work.
• Copy the groups as Markdown into your session notes, so next time you can avoid repeating them.
• End the task with a short check of what each group learned: the exit ticket generator has prompts for it. And to plan when the group sessions happen, the course schedule planner gives every lesson a date.