Grading & feedback

Lurno marks what it can automatically and hands you everything that needs a person. This article covers the two ends of that: the grading policy you set on an assessment, and what you actually do inside a single learner attempt. For finding, claiming, moderating, and bulk-releasing work across the whole organization, see The grading queue.

What this does

Objective question types are scored the moment a learner submits. Everything else — essays, open responses, uploaded files, recorded answers, image annotations — waits for a person. The grading policy decides which of those two paths an assessment takes, how many reviewers are involved, and what counts as a pass. Releasing is a separate, deliberate step, so grading and publishing results never happen by accident.

Before you start

You need permission to update the assessment to set its grading policy, and permission to record grades to mark anything.
There must be submitted work. Attempts only become gradable once a learner submits.
If an item should be scored against criteria, attach a rubric to that item in the content library first — the rubric input only appears when a rubric with criteria is attached. See Content library.

Set the grading policy

Open the assessment and go to the Grading view.

Under How submissions get marked, pick one of five methods:
- Auto-graded — the system marks every answer, instant scores.
- Mix · auto + me — multiple-choice marked automatically; written answers come to you.
- I grade everything — every submission lands in your queue.
- Two reviewers — two people grade independently and disagreements get resolved.
- Peer review — learners grade each other. Reviewers are then assigned on the peer review surface; see Peer review & portfolios.
What happens when a learner submits redraws underneath as a plain-language flow strip. For the auto and mix methods a coverage line tells you how many of the assessment’s questions can actually be auto-marked and how many still need you.
Set the Pass mark with the slider or the number field. At or above counts as passing; leave it at 0 to decide pass/fail by hand. It defaults to 70%.
Two reviewers unlocks a When reviewers disagree rule that turns several marks into one score (first pass, last pass, highest, average, unanimous, or moderator decision) and a Flag for moderation when scores differ by threshold in points. A third switch, Second review required before release, is shown but isn’t stored with the policy yet — don’t rely on it.
Choose Save grading policy. Each save creates a new version of the policy.

The grading policy is the source of truth for how an assessment is scored. The coarse “grading mode” shown in list views is descriptive only.

Grade one attempt

The same attempt drawer opens from three places: a row in the central Grading queue, a row in the queue embedded in an assessment’s Grading view, or a released attempt in the assessment’s Results view.

Responses. Each gradable item appears with the learner’s answer rendered as they gave it, plus one scoring input:
- a numeric score box against that item’s maximum,
- a pass/fail toggle for binary items,
- or a collapsible rubric block when a rubric is attached, with a score box per criterion and a running weighted total.
Items the auto-scorer already marked show an auto-scored marker and are locked until you explicitly override them.
Add a note under any response, and record a voice note for that item if you’d rather talk than type.
Write Overall feedback for the whole attempt, with its own optional voice note.
Other reviewers shows other passes’ marks. If your pass is blind, that section is withheld until you submit your own.
The footer keeps a live total and tells you how many items are still unscored.
Finish with the primary button. While items are still unscored it reads Submit grade and stays disabled; once every gradable item has a score it becomes Submit & release, which marks and publishes to the learner in one step. A separate Release button covers an attempt that was already scored elsewhere.

There is no “mark now, release later” path inside the drawer on a first pass — a completed pass releases as it submits. To hold results back for a whole cohort, set a future Scores released at on the assessment instead.

Re-opening a released attempt is read-only. If you can record grades you’ll see Edit grades; re-submitting writes a new grading record rather than overwriting the earlier one.

The assessment-scoped queue

Assessments → Grading & Review → Queue works one assessment at a time. Pick an assessment and its attempts group into Pending review, Ready to release, and Released. Opening a row here gives an attempt-level panel rather than per-item inputs: a grader pass key, a single score against the maximum, written feedback, a voice note, and Regrade / Rescore actions. Where the assessment has auto-markable questions, the panel offers the auto-marked total as a head start so you add the written-answer marks on top instead of starting from zero.

When learners see results

Releasing an attempt is what makes its score and feedback visible. On top of that, the assessment itself carries three independent release controls:

Control	Where	Effect
Scores released at	Settings → Results release	Blank releases each score as soon as it’s graded; a datetime holds every score until then
Feedback released at	Settings → Results release	The same, for written feedback — set independently of the score
When learners see correct answers	Settings → Honesty	Never / after they submit / after everyone submits, with a release datetime for the last option

Use a future scores-release datetime when you want a whole cohort to see results at the same moment. For release timing tied to a schedule, see Schedules & release windows.

Reference

Element	What it is
Grading method	Auto-graded, Mix, I grade everything, Two reviewers, or Peer review
Pass mark	The percentage at or above which an attempt counts as passing
When reviewers disagree	How several marks resolve into one score (Two reviewers only)
Flag for moderation when scores differ by	Point gap between two marks that flags the attempt for moderation
Per-item input	Numeric box, pass/fail toggle, or inline rubric — chosen per item
Auto-scored marker	The scorer already marked this item; override to change it
Item note / voice note	Feedback attached to one response
Overall feedback	A summary note (and optional voice note) on the whole attempt
Submit & release	Replaces Submit grade once every item is scored; marks and publishes in one step
Release	Publishing the score and feedback to the learner

Troubleshooting

A submission never reaches a queue. Auto-marked-only attempts are already scored and never wait for a person. Otherwise check the learner actually submitted rather than leaving a draft.

Submit is disabled. Every gradable item needs a score. The footer shows how many are still missing, right beside the running total.

Release is greyed out. The attempt has no score yet, or it’s flagged for moderation because two marks diverged past the policy’s threshold. Resolve the divergence first — see The grading queue.

I can’t see the other reviewer’s mark. Your pass is blind. Other marks stay hidden until you submit your own, which is the point — it keeps each judgement independent.

There’s no rubric input on an item. The rubric block only renders when a rubric resource with criteria is attached to that item. Attach it in the content library, then reopen the attempt.

A score is right but the learner still sees nothing. Check the assessment’s Scores released at and Feedback released at in Settings. A future datetime holds results even after you release the attempt.

I can open grading but can’t save a score. You may be able to view without holding permission to record grades. Ask someone who manages roles. See Roles & permissions.