> For the complete documentation index, see [llms.txt](https://docs.getkita.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.getkita.org/set-up-your-lesson/how-rubrics-work.md).
# How rubrics work
A rubric is the marking grid Kita shows you while you mark an assessed activity. It starts by splitting what you are judging into two axes: your criteria run down the side, and your levels run across
### Why this matters
Without a rubric, a mark comes from memory: what "good" meant for this criterion, this week, for this student, held in a way that is meant to also match what "good" meant for every other student in the class. That is hard to keep consistent and harder to defend later.
A rubric fixes the wording once, before marking starts, so:
* You are choosing from language you already wrote, not inventing a judgement on the spot.
* Two students land on the same level if their work matches the same descriptor, whoever marks them.
* A mark can be justified afterwards by pointing at the box, not by recalling the reasoning.
### Criteria, levels and cells
A rubric has up to 5 criteria and up to 10 levels. Each criterion has a title and a longer description of what it covers, and each level has a name, such as Incomplete through Mastery. Every cell where a criterion meets a level holds the wording for that combination, written by the teacher, generated by Kita, or a mix of both.
Points are set per cell, not per level, so criteria do not have to carry equal weight. A criterion you care about more can be worth 10 points at its top level while another is worth 5, and the scale can display weakest-first or strongest-first without changing what any mark means.
### Qualitative and quantitative rubrics
A rubric is either qualitative or quantitative, and the choice only affects whether marks are involved, never the wording you write.
**Qualitative rubrics** carry no points. A student sees which level they reached on each criterion, and no score is calculated.
**Quantitative rubrics** carry points in every cell. The levels you pick add up to the student's score, and that score is what gets written to the gradebook.
For a quantitative rubric, the points in the best cell of each row must add up to exactly the activity's maximum points. Kita checks the total and blocks saving until it matches, so a class can never be marked against a grid whose maximum does not exist.
### How reuse works safely
Saved rubrics sit in your own searchable list and can be attached to any assessed activity. Attaching one does not link to it: Kita copies it onto the activity. Editing the criteria, levels or wording on one activity never silently changes any other activity using the same rubric.
Kita still remembers where each copy came from. If the activity's copy has been edited since it was attached, Kita flags that it has drifted from the saved original under that name, so you know before you rely on it. Any rubric, saved or attached, can be exported as a spreadsheet to share, print or edit in bulk.
You can also build a rubric outside Kita: download the blank CSV template, fill it in, and upload it back rather than typing directly into the grid.
### How AI-assisted marking works safely
On the marking page, you open a student's submission with the rubric beside it and click one box per row. The running total and percentage update as you go, and saving writes that percentage to the student's grade.
**Mark with AI** suggests a level for each criterion. It reads the activity's description and objectives, the rubric's wording and points, the student's submitted work, and the results of any automatic tests, then proposes a box per row. A suggestion is marked with a sparkle and is only ever a suggestion: nothing is saved until you confirm it. Your own selections are marked with a tick and always take precedence over an AI suggestion. You can run it across a whole class at once, and if a student's submission has not changed since the last run, Kita reuses the earlier suggestion rather than generating a new one.
Marks are only ever released to a student once you have confirmed them, so the version a student sees is the one you decided on, not a live feed from the AI.
### What a student sees
Once marks are released, a student sees the same grid, read-only, with their level highlighted on each criterion and their score if the rubric carries points. They see your decisions and the wording behind them, never the AI's suggestions, and there is nothing on the page they can change.
### Quietly happening in the background
An activity keeps its own frozen copy of the attached rubric, so a student is always marked against the version that was in place when marking happened, even if the saved original is edited afterwards.
Nothing you explore in the rubric builder is lost by accident. Previewing a saved rubric sets your own unsaved work aside rather than overwriting it, and closing the builder with unsaved changes always asks first.
### What a rubric is not
* **Not required.** An assessed activity works without one. You can accept submissions and return written feedback with no score at all.
* **Not a live document.** Editing a saved rubric never retroactively changes marks already given on an activity using an older copy of it.
* **Not an automatic grade.** Mark with AI proposes levels. It does not save a mark, and it never releases one to a student without you confirming it first.
### Frequently asked questions
Can I use a rubric without giving marks?
Yes. Choose a qualitative rubric. Students see the level they reached on each criterion, but no score is calculated or added to the gradebook.
Does editing a saved rubric change activities that already use it?
No. Attaching a rubric copies it onto the activity. Editing the saved original afterwards does not change any copy already attached elsewhere, though Kita flags that the two have drifted apart.
Does the AI ever mark a student without a teacher?
No. Mark with AI only ever suggests a level, shown with a sparkle marker. Nothing is saved, and nothing reaches a student, until you confirm it yourself.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.getkita.org/set-up-your-lesson/how-rubrics-work.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
