Feedbacker. Share

See it in use

Marking and moderation, step by step.

Every picture here is of Feedbacker running with the made-up coursework that comes with it: no real student, mark or comment. They are made automatically by Feedbacker's own checks, so they show it as it is now.

Marking a cohort, with feedback

From the bulk download to feedback you approved.

  1. The Home screen, listing two marking workspaces, each with when it was started, how long it is kept for, its progress (0 of 7 steps done, and the next step) and a Continue button.
    Home. Each piece of work under marking or moderation, with how far it has got and what comes next.
  2. The cohort's submissions: two imported as sub-001 [STUDENT_A] and sub-002 [STUDENT_B], with their IDs, file types and any warnings, and one file in the download listed as not imported.
    Importing the cohort. Every submission in the platform's bulk download, each under a pseudonym. A file whose name carries no student ID is listed, never guessed at.
  3. Reviewing a submission's anonymised text: a warning that redaction can't see inside images, a bar chart shown where it was in the text with a Don't send tickbox, and the start of the list of replacements.
    Anonymisation. Names, IDs and contact details are replaced on your computer. Images can't be anonymised, so each is shown where it was, for you to send or keep back.
  4. The marking screen: where this submission stands, the criteria as buttons, the anonymised submission on the left, and on the right the rubric's levels for the first criterion, a mark of 85 chosen, and a comment box.
    Marking. The submission beside the rubric. The AI's suggested level, if you ask for it, is only a starting point: you choose every level and every mark.
  5. The feedback screen's guide: version 1, approved, with a line of guidance in its text box.
    The feedback guide. What each level typically needs to hear, approved by you before it is sent with every draft, so similar work gets similar feedback.
  6. Writing one submission's feedback: a criterion's draft in a text box, with buttons to draft it again, clear it and write your own, or record it, and a note of one flag to check.
    Writing the feedback. The AI drafts from your own marks and comments; you edit and record it. Each piece is checked against its mark.
  7. An AI-suggested edit to flagged feedback, with buttons to use it or not; below, the flag (Excellent is praise for first-class work, but the mark is upper second) and ways to keep the feedback with a reason.
    A flag, and a suggested edit. Praise above the mark is flagged. You can use the AI's suggested edit, change it, or keep your words with a reason.
  8. Feedback across the cohort: for one criterion, a table of each submission's level, mark and feedback, with an Edit button on each.
    Across the cohort. Each criterion's feedback side by side, grouped by your level, so you can see similar work gets similar feedback.
  9. What sub-001 will receive: an overall mark of 72.3, each criterion's mark out of 100 with its feedback, and an Approve exactly this button.
    What a student receives. Exactly the marks and feedback that will be released, for you to approve as they are.
  10. The export screen: one submission approved and another not ready yet, with what the approved one will receive below.
    Approved and exported. Only what you approved is exported: each student's feedback, and a table of marks, ready to paste into the marking platform.

Moderation

Checking a colleague's marking, with a second reader.

  1. Reviewing a sampled submission openly: the anonymised text with a quote highlighted on the left; on the right, the AI reading's suggested level, the quote it cites with a Show in the text button, and your level for the criterion.
    An open review. The AI's independent reading beside the submission. Every quote it cites can be shown in the text.
  2. A blind review after the reveal: each criterion's judged level, the submission's text, and a revision recorded for the first criterion beside the first judgement.
    A blind review. Your own judgements first; then the original marking and the AI's reading are revealed, and any revision is recorded beside your first.
  3. The moderation overview: how far it has got, and a table of the two sampled submissions with their band, each step's status, what has been judged and the verdict.
    The overview. How far each sampled submission has got, and where you, the original marker and the AI agree.
  4. The moderation's export screen: approved, with nothing changed since, the approved summary to open, and the moderator's overall comment.
    The moderation record. Approved by you, and exported as a summary ready for the moderation form, in Word and Markdown.

Next

Try Feedbacker with the made-up coursework.

Free, open source, and it runs on your own computer. Nothing about real students goes anywhere until you approve it.