Describe the assignment and Joy drafts a grading rubric with clear criteria, four performance levels, and point weights, so two graders reach the same score and every learner knows what good looks like.
Build a grading rubric for a [written case-study analysis] assignment with [4] performance levels.
An illustrative example. Your output will reflect your own sources, data, and prompts.
Here's a rubric for the case-study analysis, built from the assignment brief and the course objectives in your Knowledge Center. Five criteria, four performance levels each, weighted to 100 points with a pass bar at 70.
Each row is scored at one of four levels, and the points for that criterion scale with the level: full points at Exemplary, about three-quarters at Proficient, half at Developing, and a quarter or less at Beginning. Add the five criterion scores for the final grade.
| Criterion | Exemplary | Proficient | Developing | Beginning | Points |
|---|---|---|---|---|---|
| Problem analysis | Frames the core problem precisely and surfaces the underlying causes and trade-offs. | Identifies the main problem correctly with mostly sound reasoning. | Names a surface problem but misses key drivers or context. | Misreads or simply restates the prompt with no real analysis. | 30 |
| Use of evidence | Every claim is backed by specific, well-chosen evidence from the case. | Most claims are supported; a few assertions go unbacked. | Evidence is thin, generic, or only loosely tied to the claims. | Assertions stand with little or no evidence from the case. | 25 |
| Recommendations | Specific, actionable recommendations that follow directly from the analysis. | Reasonable recommendations, if under-specified or partly unsupported. | Vague or generic recommendations weakly linked to the analysis. | Missing, or disconnected from the problem identified. | 20 |
| Structure & clarity | Logical flow and clear signposting; the reader follows the argument effortlessly. | Organized and readable with minor lapses in flow. | Structure is uneven; the argument is hard to follow in places. | Disorganized, with no discernible line of reasoning. | 15 |
| Mechanics & citations | Clean prose; sources cited correctly and consistently. | Minor errors that don't impede reading; citations mostly complete. | Frequent errors or inconsistent, incomplete citations. | Errors throughout; citations missing or incorrect. | 10 |
A submission passes at 70 of 100. Problem analysis and use of evidence carry more than half the weight between them, so a strong argument backed by the case matters most, while mechanics and citations round out the score without dominating it.
Want me to turn this into a student-facing checklist so learners can self-check before submitting, or add a grader notes column with the evidence to look for at each level?
Describe the assignment and Joy drafts a rubric: the criteria that matter, four named performance levels with a plain descriptor for each, and point weights that add up to a total. It gives every grader the same yardstick and shows learners exactly what earns each score.
Tell Joy what the assignment is and paste the brief or the learning objectives from your Knowledge Center if you have them. No template required.
Say how many performance levels you want and any criteria you already know matter. Joy drafts the full rubric, or start from its /draft command.
Check that the criteria cover what you actually assess and that each level descriptor is distinct. Adjust weights, rename a level, or split a criterion in a follow-up.
Copy the finished rubric into your LMS, a shared grading sheet, or the assignment doc. Joy drafts it; you paste it where your graders work.
Save this ask as a custom command on the assistant your team already uses, so anyone can run it in one step.
Each performance level gets a one-line descriptor written so two graders read the same work the same way.
Criteria carry point weights that sum to a clear total, with a pass threshold you can set, so scores are defensible.
Point Joy at the assignment brief and learning objectives and the criteria map to what the work is actually meant to demonstrate.
Rename a level, tighten a descriptor, reweight a criterion, or add a row with a single follow-up request.
Turn the same rubric into a plain checklist learners can self-check against before they submit.
Add a column for the evidence a grader should look for at each level, so scoring is faster and more consistent.
Swap the criteria for delivery, content, and Q&A to grade a live presentation instead of written work.
Ask for a simpler three-level scale or a finer five-level scale and Joy rebalances the descriptors and points.
Describe the assignment and say how many performance levels you want. Joy drafts a rubric with criteria rows, a plain descriptor for each level, and point weights that sum to a total. It draws on the assignment brief and objectives in your Knowledge Center if you connect them, and you refine any criterion or descriptor with a follow-up.
Yes. Ask for a three-level or five-level scale and Joy rebalances the descriptors and points. You can reweight any criterion, rename a level, split or merge rows, or move the pass threshold, all in a follow-up in the same chat.
No. Joy drafts the rubric as text in the chat. Once it reads right, you copy it into your LMS, a shared grading sheet, or the assignment doc where your team grades. Joy writes it; you place it where you work.
That's the point of the level descriptors. When each level has a distinct, plain description of what the work shows, two graders reading the same submission are far more likely to land on the same score, and a learner can see exactly what earned their grade.
Yes. Ask Joy to turn the rubric into a student-facing checklist and it rewrites the criteria as things a learner can self-check against before submitting, so they know what good looks like ahead of grading.
Join the waitlist and be first to try this workflow when JoySuite launches.