Edit a live assessment

Why submitted results never change

A candidate's results are computed from that attempt's stored per-response scores: the automatic score recorded when they answered, and the rubric scores your evaluators entered. Editing the assessment's structure never rewrites those stored scores, so adding, editing, reordering, or removing items does not change a candidate who already submitted.

Items you add after a candidate submitted are simply skipped in that candidate's rollup — a missing response is never treated as a zero, so new questions cannot silently drag earlier candidates down. The builder shows a notice on live assessments confirming this guarantee while you edit.

What weight and points changes do affect

There is one honest exception to "nothing changes": aggregation settings. Role-fit is a weighted rollup computed from stored scores using the current section weights, skill-mapping weights, and item points. Raw scores stay frozen, but if you change a section's scoring weight or an item's points on a live assessment, the role-fit percentages are recalculated with the new weighting — for every candidate, including those who already submitted.

This is by design: everyone is always measured by the same current rules, so candidates stay comparable. But it means weight changes mid-cohort will shift the leaderboard you have already been looking at. On blueprint-linked assessments, recorded decisions also keep a role-fit snapshot from the moment each decision was made, so your decision record stays defensible.

• Item content, order, and new items: no effect on submitted candidates.
• Section weights, skill weights, item points: role-fit is re-aggregated for everyone.
• Answer-key edits apply to future submissions only; earlier responses keep their recorded score.
• On blueprint-linked assessments, recorded decisions keep a role-fit snapshot from decision time.

Candidates who are mid-attempt or yet to start

Candidates who start after your edit take the updated assessment. A candidate who is mid-attempt continues on the version they loaded; if they resume or reload, they receive the updated structure, and every answer they already saved stays attached to its item.

For edits that change what is being measured — swapping out a task, changing an answer key, re-weighting sections — think about comparability before editing mid-cohort: candidates before and after the change will have answered slightly different assessments. For a substantive redesign, the cleaner path is to duplicate the assessment into a new draft and run the next cohort on the new version.

When editing is locked

Once an assessment is closed or archived, its structure locks and its results are final — the builder tells you so and offers duplication instead. Duplicate it to a new draft to make changes; the duplicate carries the structure, weights, and rubric snapshots over.

Editing also requires structure-edit permission in your workspace. If you can see the builder but not change it, a notice explains that your role does not include structural edits.