Grading submissions

Review Assignments is the queue: everything awaiting a mark across every course you teach. It is the one screen to work from. Grading from inside each course means carrying the queue in your head.

The queue

Two tabs, Unreviewed and Reviewed.

Each row shows the learner, the assignment, when it arrived, how many attempts they have made, and the reviewer if someone has picked it up. That last column is what stops two instructors marking the same submission twice.

Assign reviewer claims a submission, or hands it to a colleague. On a course with several instructors, claiming before you start is worth the click.

Filter by course or cohort when you want to grade one batch in one sitting. Marks are measurably more consistent within a sitting than across several.

Marking

Opening a submission puts the work beside the marking form.

What the learner submitted: a repository link, and optionally a live URL, a file, their own comments and a difficulty rating. All the links open in a new tab.

What you fill in: a score out of the assignment's maximum, and written feedback. The pass mark comes from the assignment, and whether the mark passes is shown as you type.

There are quick templates for common openings — Excellent work, Good effort, Great UI/UX design, Needs revision. They are a starting sentence, not a review. A template on its own tells a learner nothing they could act on.

On feedback

The mark tells a learner where they landed. Only the comment tells them what to do differently, and that is the part they act on.

Three things that make feedback land:

Point at the thing. "The submit handler in App.jsx runs before validation" beats "validation issues".

Say what you would do instead. A criticism with no alternative is a verdict.

Lead with the most important one. Eleven equally-weighted notes get read as noise; two ranked ones get acted on.

If the assignment has a rubric, mark against it. It is why it exists, and it is what a learner will point at if they disagree with you.

The AI pre-review

Where your organisation has it switched on, a draft review is prepared for each submission and shown in its own panel.

You have three choices:

• Use it — copies the text into your feedback box, where you edit it.
• Reject it — discards this draft.
• Ignore it — write your own.

Nothing generated reaches a learner without a person sending it. The panel is a draft for you, not a review for them.

It is decent at noticing what is present or absent in a submission and poor at judging whether that mattered for this learner at this point in the course. Use it for the mechanical sweep and keep the judgement.

Returning work

Grades are returned per submission as you finish each one, and the learner is notified.

To hold a whole assignment back instead — so nobody sees their mark before the rest of the cohort — grade the batch and release it together. Until it is released, marks are visible to staff only.

Asking for another version

Setting a submission to Review again tells the learner you would rather see the work improved than mark it as it stands. It reopens the assignment for resubmission.

The previous mark and comments are kept on the record rather than overwritten, so the history of what was returned stays intact and you can see, on the next attempt, whether your points were addressed.

Used well, this is the single most useful control on the page. A learner who fixes the thing and resubmits has learned something; a learner who got 48% and a paragraph has learned considerably less.

Bonus marks

Beyond the assignment scheme, you can award bonus points on a learner's cohort enrolment, with a reason. Bonus sits on top of the 100 the normal categories share out — it is extra credit, not a re-weighting.

Courses can cap the total bonus available. Where the course's grade scale allows it, bonus can carry a final mark past 100.

Always fill in the reason. It is the only record of why someone's total does not follow from their marks, and six months later it is the only thing that can answer the question.

How the course mark is assembled

Worth understanding, because learners will ask.

1. Categories and weights. Work is grouped into categories, each worth a share of 100. The default is Assignments 45, Quizzes 30, Attendance 25. Each category is scored as a percentage of what was available in it, then multiplied by its weight.

2. Bonus is added on top.

3. A grade scale maps the percentage to a grade — A/B/C, Distinction/Merit, Pass/Fail, whatever your organisation uses.

The categories and weights for a course are on its Grading tab, and the weights must add up to 100. Set them before the first mark is given. Changing them mid-run re-weights everything already awarded.

The scheme is snapshotted per offering. A cohort running a course keeps the scheme as it was when the offering was created, so editing the course's scheme does not silently rewrite a cohort already half way through.

The scales themselves are set by administrators under Grading Bands. See Grading scales.

Attempts, and which counts

A learner who takes a course twice has two attempts, each with its own mark. Their record takes the best. Attempts are not averaged and a later worse one does not replace an earlier better one.

When a mark shows as a dash

A dash means the maximum for that item is not known, so a percentage cannot be computed. It is not a zero.

That is a deliberate choice: being silent about a mark that cannot be worked out is recoverable, and being confidently wrong about a learner's grade is not. A dash where you expect a number usually means an assignment was created with no maximum score.