DEV Community

ToolBrewLab
ToolBrewLab

Posted on Edited on

Split the class without scrambling the seating plan

Every teacher has a seating plan that took real thought. The talkative pair sits apart, the quiet one sits near a confident partner, the kid who needs the front row is in the front row. Then group work starts and the random name generator does what random generators do: it shuffles all of it away. The groups it makes are fair in the statistical sense and useless in the classroom sense, because the order that mattered is gone.

That is the exact problem this tool is built around. I use the Split Group in Order with Rotation for it now. You paste your names in the order that matters, seating order, registration order, rota order, and the order is never shuffled. What you type is exactly what the tool keeps. Then you choose how to split: by group size, say groups of 4, or by number of groups, say 4 groups. Set how many weeks the schedule should cover and the rotation step, and it builds the groups plus the full rotation table.

Most group tools treat order as something to randomise away. This one treats it as the input. If your seating plan is doing real work, keep it. The order is the data. A seating plan records who sits next to whom, a registration list records the order people signed up, and a rota can carry seniority or fairness. Scramble the names and all of that disappears for good. Splitting in order keeps it intact and still gives you fair, balanced groups.

The remainder rule is the detail that matters

Thirteen names into groups of 4 gives 4, 3, 3, 3. The larger group goes first instead of leaving one group short at the end or one sad group of one. That is the remainder rule, and it is the detail that shows someone actually thought about this. Bigger groups always come first, and the tool never leaves anyone in a group of 1 without asking you first.

A plain split of 13 into groups of 4 leaves one person on their own: 4, 4, 4, 1. Anyone who has taught a class knows what that looks like. One student on their own, doing group work alone, while the rest of the room works in fours. It is not a group. It is an embarrassment, and it happens because the splitter did not think about the remainder. This one puts the spare names into the earliest groups instead, so you get 4, 3, 3, 3, and nobody works alone.

Pick group size when the size itself matters, tables that seat 4, for example. Pick number of groups when the count matters, 3 breakout rooms, say. Either way, your order is kept and the same bigger-groups-first remainder rule applies. That choice is the whole split, and it takes about ten seconds to make.

A worked example with real names

Say you have got a class of 13 and need groups of 4, rotated over 3 weeks. Paste the 13 names in order, choose by group size with a size of 4, set 3 weeks and a rotation step of 1. The tool hands back groups of 4, 3, 3, 3, the remainder rule doing its job. Nobody ends up alone, and every name is still sitting exactly where you put it.

Week 1 gives you Group 1 with Ava, Ben, Cleo and Dan, Group 2 with Eli, Fay and Gus, Group 3 with Hal, Ivy and Jon, and Group 4 with Kim, Leo and Mia. Week 2 shifts the whole list forward one place and splits it again. Ben, Cleo, Dan and Eli land in Group 1, while Ava wraps round to the end of Group 4. Week 3 shifts two places from the original order. Same names, same group sizes, but a different mix, and you can repeat or extend the schedule as far as you like.

A step of 1 nudges everyone one place forward each week, so nobody sits in the same seat or the same group all term. Over a few weeks every student works with different classmates, which is the whole point of rotating in the first place. Each week the whole list moves forward by the step, wrapping round at the end, then gets split again. With a step of 1, whoever is at the top sits second in week 2 and third in week 3. A bigger step, like 2, jumps further each week for faster mixing.

A six-week project, planned in one paste

Say you have 27 students and you want groups of 4 for a six-week project. You paste the 27 names in seating order, pick by group size, set 6 weeks and a step of 1. The first week gives you six groups of 4 and one of 3, and every week after that the groups shift by one place. Copy the result into a document or print it, and the seating plan you spent an evening on survives the whole term.

That is the use case I keep coming back to, because it is the one where the random generator does the most damage. A six-week project is long enough for the seating plan to matter and long enough for fixed groups to go stale. Fixed groups go stale quickly. The same people always end up together, the same pairs do all the talking, and the quiet ones never meet anyone new. The rotation table shifts the list by your step each week, so the mix stays fresh without any extra planning. You set it once, on the first day, and the whole term is planned.

There is no shuffle button and no randomisation anywhere in the tool. Your names stay in the exact order you typed them, in week 1 and in every rotation week. That sounds like a small thing until you have watched a randomiser eat a seating plan you spent an evening building. Once it is shuffled, that information is gone. There is no undo.

Three mistakes I kept making

Using a random splitter was the first. Randomisers look fair, but they wipe out the seating order, registration order or seniority your list was built on. Once it is shuffled, that information is gone, and there is no undo. Ignoring the remainder was the second. A plain split of 13 into groups of 4 leaves one person on their own, 4, 4, 4, 1, and this tool puts the bigger groups first instead, so you get 4, 3, 3, 3. Repeating the same groups every week was the third. Fixed groups go stale quickly, the same people always end up together, and the rotation table fixes that by shifting the list by your step each week, so the mix stays fresh without any extra planning.

The tool does one job and it does it properly. Hit Split groups to see your groups and the full rotation table. Copy result drops the schedule into a document, and Print gives you a paper copy. Clear clears the result tables. That is the whole interface, and it is enough, because the job is not complicated. The complicated part was the order, and the order is safe now.

Questions I get asked

Why does order matter more than random? Because the order is the data. A seating plan records who sits next to whom, a registration list records the order people signed up, and a rota can carry seniority or fairness. Scramble the names and all of that disappears for good. Splitting in order keeps it intact and still gives you fair, balanced groups.

How does the remainder rule work? When the names do not divide evenly, the spare names go to the earliest groups. Thirteen names in groups of 4 become 4, 3, 3, 3. Bigger groups always come first, and the tool never creates a group of 1 without checking with you first.

How does the rotation step work? Each week the whole list moves forward by the step, wrapping round at the end, then gets split again. With a step of 1, whoever is at the top sits second in week 2 and third in week 3. A bigger step, like 2, jumps further each week for faster mixing.

Most group tools treat order as something to randomise away. This one treats it as the input. If your seating plan is doing real work, keep it.

Top comments (0)