← IndexEntry № 213·education

Build a Standards-Aligned Analytic Grading Rubric

Generates a weighted analytic rubric with observable performance-level descriptors tied to your learning objectives.

Optimized for
ChatGPTClaude
§ When to use this

A grading rubric is only useful if students can actually use it to predict their grade and improve their work. Vague rubrics that say 'good organization' versus 'poor organization' leave everyone guessing and make grading feel arbitrary. This prompt builds an analytic rubric, meaning each criterion is scored separately so a student who nails their argument but stumbles on mechanics gets accurate, targeted feedback rather than one fuzzy holistic number. The key is observable descriptors: instead of 'strong evidence,' the rubric specifies what strong evidence looks like at each level. Because you feed in your own learning objectives, the criteria stay aligned to what you actually taught, which matters for fairness and for defending a grade to a parent or administrator. Use it to draft a first version in seconds, then tweak the weights and wording to fit your class. It works for essays, lab reports, presentations, projects, and portfolios.

§ The Prompt— fill in the fields, then copy or open in a tool
§ Customize0/5 fields filled
your prompt — fill the fields above
You are an assessment design specialist. Build an analytic grading rubric for a [ASSIGNMENT TYPE] in [SUBJECT] at the [GRADE LEVEL] level, worth [TOTAL POINTS] points. Identify 4-5 distinct criteria (e.g., thesis, evidence, organization, mechanics) drawn from these learning objectives: [LEARNING OBJECTIVES]. For each criterion, write 4 performance levels (Exemplary, Proficient, Developing, Beginning) with concrete, observable descriptors a student could self-check against. Weight criteria by importance and show the point split. Use plain, non-judgmental language and avoid vague words like 'good' or 'poor'. Output as a clean table.
Open with your prompt →ChatGPTClaudeSends your filled-in prompt straight into a new chat.
§ Example Output

What you can expect back

| Criterion (weight) | Exemplary | Proficient | Developing | Beginning |
|---|---|---|---|---|
| Thesis (25 pts) | States a debatable claim that previews the argument | States a clear claim | Claim is present but broad or factual | No identifiable claim |
| Evidence (30 pts) | 2+ relevant cited sources, fully analyzed | 2 cited sources, some analysis | 1 source or weak analysis | No cited evidence |
| Organization (25 pts) | Logical flow with smooth transitions | Clear structure, basic transitions | Some sequencing issues | Ideas feel random |
| Mechanics (20 pts) | Virtually error-free | Minor errors, no meaning lost | Errors occasionally distract | Errors impede reading |
Point split totals 100. Each cell is observable so students can self-assess before submitting.

Illustrative example — your results will vary by tool and inputs.

§ Pro Tips

Get sharper results

  • 01Paste a sample student response and ask the model to score it against the rubric to pressure-test whether the levels are distinguishable.
  • 02Ask for a student-friendly version that rewrites each descriptor in first person ('I stated a debatable claim') for self-assessment.
  • 03Keep weights as round numbers that sum to your total so mental math during grading stays fast.
  • 04Have the model generate one 'anchor' example per level so your grading stays consistent across a stack of papers.
§ Variations

Adapt it for your case

Single-point rubric

Ask for a single-point rubric: one column of proficiency descriptors with blank space on either side for 'concerns' and 'exceeds', which is faster to grade with.

Holistic version

Request a 4-level holistic rubric instead, useful for quick formative checks or timed in-class writing.

Standards-coded

Add 'tag each criterion with the matching state standard code' to map the rubric directly to your standards documentation.

Best For — Roles
Use For — Tasks
Tags#rubric#assessment#grading
§ FAQ

Common questions

What's the difference between analytic and holistic rubrics?

Analytic rubrics score each criterion separately so feedback is targeted; holistic rubrics give one overall score. Use analytic when you want students to know exactly where to improve.

Will the point weights always add up correctly?

Usually, but always check the sum yourself. Ask the model to show the running total, and adjust any criterion if the split doesn't match your total points.

Can I reuse one rubric across assignments?

Yes, for the same skill set. If the assignment targets different objectives, regenerate it with the new objectives so the criteria stay aligned to what you taught.

§ Related Entries

You may also need