Rubric Generator

{{ rubricTypeLabel }} {{ activeCriteria.length }} criteria {{ levelCountLabel }} {{ weightBadge }} {{ auditBadge }}

Assignment title:

Keep this short enough for table headers and exported notes.

Learner level:

Choose the audience that will read the rubric before completing the work.

Learning goals or assignment brief:

Use one or two concise outcomes; the audit flags blanks because rubrics should connect back to the task.

Rubric type:

Analytic is the default when students need criterion-by-criterion feedback.

Performance scale:

Four levels is a common classroom balance between focus and useful progress signals.

Total points:

Use the assignment total, such as 20, 25, 50, or 100 points.

pts

Rubric criteria:

Weights are normalized automatically, but the audit still flags totals that are far from 100.

Criterion label:

Examples: Content accuracy, Evidence and explanation, Visual organization.

Observable evidence:

Write what the work should show when this criterion is strong.

Weight:

Use percent-style values for clearer audits, such as 30 or 25.

weight

Samples replace the current assignment and criteria.

Check rubric setup

{{ error }}

Descriptor depth:

Use standard for classroom distribution unless the rubric must fit on one page.

Point rounding:

Whole points are simple for classroom grading; half points preserve more allocation detail.

Standard or course code:

Optional short code, such as NGSS MS-LS2-3, CCSS W.8.1, or Course Outcome 2.

Revision cues:

Leave on when students will use the rubric for self-checks or revisions.

{{ header }}	Copy
{{ cell.title }} {{ cell.meta }} {{ cell.text }}

Audit item	Status	Evidence	Next action	Copy
{{ row.item }}	{{ row.status }}	{{ row.evidence }}	{{ row.action }}

{{ markdownOutput }}

{{ jsonOutput }}

Embed:

Customize

Include current inputs

Size

Advanced

Width

Height

Aspect ratio

Max height

Collapsible embed

Allow fullscreen

Referrer policy

Sandbox tokens

A rubric earns its value before the first paper, project, lab report, presentation, or portfolio is graded. It turns a vague expectation such as "good explanation" into visible criteria, performance levels, and descriptions that students and reviewers can point to. When the language is specific enough, students know what kind of evidence to produce and graders have a shared reference for feedback.

The hard design choice is deciding what belongs in the grid. Criteria should name the qualities that matter in the work, not every task step the learner followed. Performance levels should describe meaningful differences in evidence, accuracy, reasoning, organization, or communication. Descriptors should use observable language so a student can compare the work with the expectation without guessing what the grader meant.

Criteria: The distinct qualities being assessed, such as evidence, accuracy, reasoning, organization, or communication.
Performance levels: The scale positions that describe how strong, partial, developing, or missing the evidence is.
Descriptors: The observable wording inside the rubric cells that helps students and graders recognize the difference between adjacent levels.

Rubric matrix anatomy showing criteria, performance levels, weights, and review checks

Rubrics also make tradeoffs visible. A concise rubric is easier to read during drafting, but it may hide the difference between adjacent levels. A detailed analytic rubric gives more feedback, but too many rows can make students chase the table instead of focusing on the assignment goal. Weighting adds useful emphasis when one criterion is more important than another, yet uneven weights need a clear reason because they shape the final score.

Common rubric types and when they fit
Rubric type	Best fit	Main caution
Analytic	Feedback by criterion, weighted scoring, and assignments with several distinct skills.	Too many criteria or vague descriptors can make the rubric slow to use and hard to trust.
Single-point	Formative feedback where the expected standard is clear and comments explain growth or extension.	Graders must add useful notes because the rubric does not prewrite every possible level.
Holistic	Whole-work judgment, fast scoring, presentations, portfolios, or creative work with integrated qualities.	Students receive less detail about which part of the work most affected the score.

Rubrics can improve consistency, but they do not remove professional judgment. Two graders may still interpret a descriptor differently, especially when they have not discussed sample work. Testing the rubric against a strong sample, a borderline sample, and a weak sample often reveals duplicate criteria, hidden expectations, or level descriptions that change tone instead of describing real differences.

A rubric draft should therefore be treated as an assessment plan, not just a scoring sheet. It should align with the assignment goal, use language the learners can understand, respect accommodations and local grading policy, and leave room for disciplinary expectations that no generic draft can know in advance.

How to Use This Tool:

Start with the task and learning goal, then adjust the criteria before treating the generated rubric as a reviewable draft.

Enter Assignment title and Learning goals or assignment brief. The draft works best when the goal states what students should show, not just the topic they are studying.
Choose Learner level. This changes descriptor wording for elementary, middle school, high school, higher education, or professional training readers without changing the point calculation.
Set Rubric type. Use Analytic criterion matrix for detailed criterion feedback, Single-point rubric for proficiency-centered growth notes, or Holistic whole-work scale when one overall band is enough.
For analytic or holistic rubrics, pick a Performance scale with 3, 4, or 5 levels. Set Total points between 1 and 1000 so the point allocation matches the assignment gradebook value.
Edit each Rubric criteria row. Give every row a distinct Criterion label, a concrete Observable evidence cue, and a Weight that reflects the emphasis you want in the score.
Use Essay sample or Lab sample only when you want to replace the current draft with a starting example. Use Balance weights when equal emphasis is appropriate, then reopen the criteria and adjust any row that should count more.
Open Advanced for Descriptor depth, Point rounding, Standard or course code, and Revision cues. If students will use the rubric before submitting work, revision cues usually make the descriptors more useful.
Fix any Check rubric setup alert before relying on the result. Then review Rubric Matrix, Point Allocation, and Design Audit; use Rubric Markdown or JSON only after the wording and audit checks make sense.

Interpreting Results:

Rubric Matrix is the main draft to read like a student. In an analytic rubric, each criterion shows its evidence focus, entered weight, normalized share, point value, and level descriptors. In a single-point rubric, the middle column states the expected evidence while the growth and extension columns support individualized feedback. In a holistic rubric, each performance level describes the whole work and shows a score range.

Point Allocation explains how weights become points. A clean 100-percent total is easier to explain, but positive weights are still normalized when the total is not exactly 100. Design Audit is the quality check to read before sharing the draft, because it flags setup problems that can make the rubric misleading even when the matrix looks polished.

Rubric result signals and practical checks
Result signal	Meaning	Check before use
`100% weights`	Entered criterion weights add to about 100.	Confirm the largest shares match the assignment priorities.
`Equal weights`	No positive weights were entered, so every criterion receives the same share.	Use equal allocation only when each criterion should carry the same scoring value.
`Review`	The draft can be used after a judgment check.	Read the evidence note and decide whether the issue affects students or grading.
`Adjust`	A setup problem is likely to affect trust.	Fix duplicate labels, missing criteria, weak evidence cues, or other flagged items first.

A ready audit does not prove that the rubric is fair, aligned with policy, or calibrated for the course. Compare the matrix with sample work and reread adjacent descriptors for the same criterion; students should be able to see the real difference between one level and the next.

Technical Details:

Rubric scoring is criterion-referenced. The work is judged against stated expectations instead of being ranked against other learners. In an analytic rubric, each criterion contributes its own score or feedback. In a single-point rubric, the expected standard is central and comments describe growth or extension. In a holistic rubric, one level describes the overall performance across the work.

The scoring math separates entered emphasis from displayed points. Weights act as relative emphasis values. They can be written like percentages, but the calculation first converts them into shares of the total entered weight. Rounding changes the displayed criterion points, not the underlying idea that a larger share should receive more of the assignment total.

Formula Core:

For criterion i, the normalized share is the criterion weight divided by the sum of all active criterion weights. The raw point allocation is that share multiplied by total assignment points.

\begin{array}{lcl} s_{i} & = & \frac{w_{i}}{\sum w_{j}} \\ p_{i} & = & T \times s_{i} \end{array}

Here w_i is the entered weight for one criterion, s_i is its normalized share, T is Total points, and p_i is raw criterion points before display rounding. If all entered weights are zero, each criterion receives an equal share of 1 / criterion count.

With a 50-point lab rubric and weights of 20, 25, 30, 15, and 10, the total weight is 100. The 30-weight Analysis criterion receives a 0.30 share, so its raw allocation is 50 x 0.30 = 15 points. If the same rubric used weights of 2, 2.5, 3, 1.5, and 1, the displayed shares would be the same because the relative emphasis did not change.

Rule Core:

Rubric generation rules and review effects
Area	Rule	Visible effect
Rubric structure	Analytic rubrics use criterion rows and level columns; single-point rubrics use growth, meets, and extension columns; holistic rubrics use performance-level rows.	`Rubric Matrix` changes shape and descriptor purpose.
Performance scale	Non-single-point rubrics use 3, 4, or 5 performance levels.	Matrix headers and the level badge reflect the active level count.
Point rounding	Raw criterion points can be displayed as whole, half, or tenth points.	`Point Allocation` shows raw shares and displayed point values.
Descriptor depth	Concise, standard, and detailed modes control how much descriptor text appears.	Detailed and revision-enabled drafts can add next-step language for lower-performance or growth descriptions.
Design checks	The audit reviews criteria count, weight total, duplicate labels, evidence cues, level progression, assignment alignment, revision cues, and neutral language.	`Design Audit` marks rows as ready, review, or adjust.

The technical rules create a consistent first draft from the entered criteria and settings. They do not establish inter-rater reliability by themselves. Reliability improves when graders discuss descriptors, compare sample work, and agree on what counts as evidence for each criterion and level.

Limitations:

A generated rubric is a starting draft for human review. It can expose structure, weights, descriptors, and audit warnings, but it cannot know every course policy, accommodation, disciplinary convention, or grading standard that applies to a real assignment.

Review descriptor wording for fairness, clarity, and age-appropriate language before sharing with learners.
Calibrate the rubric with sample work when more than one grader will use it.
Check that criterion weights reflect the learning goal rather than convenience or habit.
Do not treat a ready audit as proof that the rubric satisfies accreditation, program, or legal requirements.

Worked Examples:

Middle school science poster

A 100-point Analytic criterion matrix for Ecosystem food web poster uses four criteria weighted 35, 20, 25, and 20. Point Allocation should show 35, 20, 25, and 20 displayed points, and Design Audit should mark alignment and evidence checks ready when the learning goal and criterion cues are specific.

Lab report with half-point rounding

A 50-point lab report rubric uses five criteria weighted 20, 25, 30, 15, and 10 with Point rounding set to Half points. Point Allocation should display 10, 12.5, 15, 7.5, and 5 points. The 15-point Analysis row correctly receives the largest share.

Single-point revision draft

An essay draft can keep criteria such as Claim and focus, Evidence selection, and Reasoning and analysis while switching to Single-point rubric. Rubric Matrix then centers Meets standard and leaves growth and extension spaces for comments, which is useful when students will revise after feedback.

Duplicate criteria before export

If two rows are both labeled Evidence, Design Audit marks Distinct dimensions as Adjust. Rename one row to a clearer dimension, such as Source integration or Reasoning and analysis, then reread the descriptors in Rubric Matrix before using Rubric Markdown or JSON.

FAQ:

Which rubric type should I choose first?

Choose Analytic criterion matrix when students need feedback by criterion. Choose Single-point rubric when the expected standard is the main reference and the grader will add comments. Choose Holistic whole-work scale when one overall performance judgment is enough.

Do weights have to total exactly 100?

No. Positive weights are normalized before points are calculated. A total near 100 is still easier to explain, and Design Audit asks for review when the entered total may confuse graders or students.

Why did the result disappear after I changed an input?

The generator hides the draft when Check rubric setup has required fixes. Add an assignment title, keep at least one criterion, use a positive Total points value, and keep at least three performance levels for analytic or holistic rubrics.

Do revision cues change the score?

No. Revision cues change descriptor wording only. The score allocation still comes from Total points, criterion weights, normalized shares, and the selected point rounding.

Can I share the generated descriptors without editing them?

Treat the descriptors as a draft. Read them against the assignment prompt, sample work, and local grading expectations, then edit any language that is too vague, too harsh, duplicated, or not observable in student work.

Glossary:

Analytic rubric: A rubric that separates criteria and usually gives each criterion its own performance descriptors and score contribution.
Single-point rubric: A rubric centered on the expected standard, with room to describe growth and extension evidence.
Holistic rubric: A rubric that describes whole-work performance levels rather than scoring every criterion separately.
Criterion: A distinct quality of the work being assessed, such as evidence, accuracy, organization, or communication.
Descriptor: The text explaining what performance looks like for a criterion and level.
Normalized share: A criterion's fraction of the score after its entered weight is divided by the total entered weight.
Inter-rater reliability: The degree to which different graders apply the rubric consistently to the same work.

References:

Rubrics, Teaching@UW.
Creating and Using Rubrics, Carnegie Mellon University Eberly Center.
Rubrics, University of Illinois Chicago Center for the Advancement of Teaching Excellence.
Creating Rubrics, DePaul University Teaching Commons.