How to make balanced student groups

Choose the attribute before choosing the control

• Write down why this activity benefits from complementary knowledge, comparable preparation, or a spread of roles.
• Prefer observable, temporary information such as a chosen topic, completed prerequisite, or self-selected role.
• Use categorical balance for category distributions and numeric attribute sums for totals such as an independently justified rating.
• Avoid diagnoses, disability, behavior reputation, family circumstances, or informal staff judgments.
• If the attribute would not change a decision, do not collect it.

Configure and review balanced groups

1. 1

   Set the class and group shape

   Confirm attendance, group count or size, and any unavoidable smaller group before adding balance goals.

2. 2

   Import one useful field

   Give the attribute a neutral name and use consistent values. The preloaded example uses a self-selected project role rather than a judgment about ability.

3. 3

   Choose the matching preference

   Use Attribute Balance for categories. Use Attribute Sum only for meaningful numeric values whose totals should be distributed.

4. 4

   Separate requirements from preferences

   Fixed assignments and must-stay rules are hard requirements. Balance is a best-effort preference and should not be described as guaranteed.

5. 5

   Generate privately

   Review group composition before showing students. Remove private columns from participant-facing exports or screenshots.

6. 6

   Evaluate the decision

   Compare the per-group counts with the arithmetic target and ask whether the attribute improved the activity rather than merely making the roster look orderly.

Trace a balance claim back to the actual groups

• The assigned count and capacity under Your groups reveal uneven group sizes immediately; the balance score does not.
• The Attribute Balance card turns an aggregate penalty into something reviewable: each deviation names the session, group, category, desired count, and actual count.
• Required fixed, together, and apart rules are solver invariants. Their cards report “No violations” in a valid result; anything else is a solver defect to report, not a compromise for the teacher to repair.
• Across several sessions, the Pair Meeting Matrix can expose repeated classmates. It cannot show who felt isolated, carried the work, or needed support; those remain human review questions.
• Assignment-only exports are suitable for sharing only after the downloaded or copied result has been checked for private attributes and reasons for constraints.

Open the relevant GroupMixer sections

These links open your current scenario. Use the guide example’s explicit handoff to create a separate copy of its synthetic classroom setup.