Rubric Generator

How to use it

1. Type a rubric title and, if you want, a class or assignment label below it.
2. Click a template to start from a worked example, or edit the default rubric in place.
3. Edit the performance level headers across the top — the label (Excellent, Proficient, etc.) and the point value for that level.
4. Edit the criterion name on the left of each row, then write a descriptor for what each level looks like in each cell.
5. Use + Criterion to add a row and + Level to add a column. The small × button on a header removes that row or column.
6. Click Print / Save as PDF. The print dialog hides everything except the rubric, the title, and the AnchorKite credit line.

Use Copy as Markdown to paste the rubric into a Google Doc, a learning-management system, or a syllabus file. Use Reset to clear the saved rubric and return to the starter template.

Analytic vs. holistic rubrics

An analytic rubric scores several dimensions of the work separately: a five-paragraph essay might get a 4 on thesis, a 3 on evidence, a 2 on organization, and a 3 on mechanics. A holistic rubric collapses all of that into a single overall judgment per performance level: this paper, taken as a whole, is a 3.

Analytic rubrics produce better feedback. A student who lost points sees exactly which dimension cost them — the thesis was unclear, or the evidence was thin, or the organization wandered — and can target the next draft at the weakest column. Holistic rubrics are faster to grade and useful for high-stakes summative judgments where students will not revise, but they communicate less about how to improve.

This tool produces analytic rubrics by default. If you need a holistic rubric, keep a single criterion row and write the four descriptors as overall-quality bands. Or use the tool to draft an analytic rubric for your own grading, then collapse the columns mentally when assigning a final letter.

Writing descriptors that actually help

The hardest part of any rubric is writing the cell text — the descriptor that says what "Excellent" or "Developing" looks like on the dimension this row is measuring. A few rules of thumb:

• Describe observable features, not value judgments. "Well-written" is not a descriptor; it is a grade in disguise. "Sentences vary in length and rhythm; transitions are explicit; word choice is precise" is a descriptor — you can point to a sentence and say whether it meets the bar.
• Use the same dimension across all levels. If "Excellent" describes thesis specificity, "Developing" should also describe thesis specificity, not switch to talking about grammar.
• Write the top and bottom first. The middle bands are harder to articulate. Defining "Excellent" and "Beginning" first makes "Proficient" and "Developing" the interpolation between them.
• Keep cell descriptors short. Two to four lines is plenty. A rubric that is dense paragraphs in every cell is unreadable while grading. Save the long explanation for the assignment prompt.
• Pilot it on real work. Grade three actual papers with the draft rubric before handing it to students. The cells you cannot match to real work are the cells that need rewriting.

How point values work

Each performance level carries a point value, shown in the column header. The default scale is 4 / 3 / 2 / 1, which makes the highest level worth the same as the number of levels. The point total at the bottom of the rubric is highest level value × number of criteria — so four criteria with a top level of 4 points gives 16 points total.

If you want some criteria to weigh more than others, change the point values in the level headers and use the column structure to set the weighting (a 4 / 3 / 2 / 1 column on one rubric versus a 10 / 7 / 4 / 1 column on another communicates different stakes). For most classroom uses, the simple equal-weight model is enough — the writing-quality benefits of a clear rubric matter more than fine-grained point engineering.

Frequently asked questions