Make contribution visible without dismantling collaboration
Group assessment may test subject knowledge, coordination, negotiation and the integration of different perspectives. A record should illuminate those achievements, not require every minute to be attributed to one person. It asks who took responsibility for an output, who shaped it, how it was checked and where the group made a collective decision.
File ownership is weak evidence on its own. One student may frame an analysis that another implements, while a third finds a defect that changes the conclusion. Conversely, a long edit history does not establish intellectual significance. Describe activity in relation to a result, decision or quality check.
Whether individual and group marks may be combined, and with what weighting, depends on the approved assessment. Publish that arrangement before work begins. This template is discipline-open and does not create an institutional regulation.
Plan outputs and interfaces before assigning names
Start with observable outputs such as a question, dataset, prototype, interpretation, report or performance. Do not fragment the project into trivial actions. The group then assigns roles, dates, handover conditions and shared review points. Responsibility becomes usable only when those elements are combined.
The planned matrix is provisional. Illness, access problems or a changed research need may alter contributions. Record the date, reason and agreement for a change. Work must not silently migrate to another member and later remain credited to the original allocation.
Do not force collective work into an individual box. For an important shared decision, note the proposal, material alternatives, decision and subsequent check. Facilitation, integration and conflict resolution may be substantive contributions, but they are not automatically more valuable than technical or conceptual work.
Use limited academic evidence, not continuous surveillance
Suitable evidence can include versioned outputs, decision notes, analysis or test logs, review comments and concise personal reflections. Complete private chats, attendance minutes and permanent screen recording are disproportionate defaults. Metadata never proves intellectual authorship in isolation.
Each member can answer three questions: Which output did I materially shape? Which collective decision did I support? Which contribution from another person did I review or integrate? Specific examples communicate more than unsupported percentages with no defined denominator.
AI assistance and external help are recorded according to the task rule. A system cannot occupy an accountable group role. Credentials, protected research material and private third-party communication should not be copied into a broadly accessible contribution record.
Separate individual evidence from the shared product
A discipline-open rubric can consider the quality of an individual's output, reasoning behind decisions, interface work, review of others and reflection. Assess the collective product through the published group criteria. Weightings and minimum requirements must be available before students begin.
Peer reports are evidence, not automatic marks. Friendship, conflict, status and the visibility of different tasks can distort them. Ask for behavioural examples, compare perspectives and allow a response. A majority vote cannot determine the performance of an unpopular group member.
An unsuccessful group decision may still reveal individual learning when someone identified a risk, explained an objection and helped correct the result. A polished final product does not prove equal contributions either. Retain both levels rather than forcing one to stand for the other.
Resolve discrepancies early and through a protected route
Use brief scheduled reviews during the project. Address a missed handover factually and revise the plan first. If disagreement remains, provide the marker with the relevant output, current matrix and the different accounts. Every affected person should have a fair opportunity to respond.
Dependency, harassment or discrimination may make direct group discussion unsafe. Publish a confidential route that does not require a student to confront another member. Restrict sensitive records to staff who need them and retain them only under the institution's applicable process.
Do not retrospectively convert a normal project problem into misconduct. Academic misconduct and contribution assessment require their respective authorised procedures and evidence standards. Record uncertainty explicitly when evidence cannot resolve the issue.
Apply a documented final cross-check
For each matrix row, ask whether the output is observable, the role corresponds to a recorded decision and at least one academic artefact supports the claimed contribution. Do not average contradictions away. Mark each account as established, disputed or not evidenced, then explain how the published criteria were applied.
A late handover may be organisationally relevant but does not alone prove poor academic quality or no participation. Equally, a highly visible presentation cannot eclipse less visible analysis, review or integration. After marking, give the group a privacy-conscious explanation of the contribution categories and process used. Keep personal conflict notes within the protected route.
Use the AI disclosure rubric where tools contributed to outputs. The reproducibility grid helps review evidence chains. Browse the higher education practice hub, and use the one foundation route, how verifiable checking works, for an evidence-oriented overview.