Turning Learning Objectives Into Rubric Criteria

Every assignment starts with a learning objective — what you want students to know or be able to do. A rubric turns that objective into something you can score: criteria with descriptors that show what meeting the objective looks like at different levels of quality. The gap between the two trips up a lot of rubrics — an objective like “students will analyze primary sources” is a fine instructional goal, but pasted in verbatim as a criterion it doesn’t tell a grader, or a student, what analysis has to contain on the page. This guide is a repeatable way to close that gap: read the objective’s verb, decide what evidence would count, write it in language a student can use, and check it across a few subjects. New to rubrics? Start with How to Write a Rubric first; this guide zooms in on the step most people find hardest.

Objectives vs. criteria

A learning objective describes an outcome for instruction — one sentence, often shared across a whole unit (“students will evaluate the credibility of sources”). It’s written for planning, not for grading a specific paper on a specific day. A rubric criterion is narrower: a named dimension of one task — “Source Evaluation” — with descriptors for what a grader will actually see in the work at each level. A criterion has to survive contact with a real stack of student papers; an objective doesn’t.

The most common failure is skipping the translation step — pasting the objective in as the criterion name and expecting descriptors to write themselves. “Analyzes primary sources” as a heading, scored “does this well / somewhat / not at all,” measures nothing; it restates the question instead of answering it. Wiggins and McTighe’s backward-design model, from Understanding by Design, is useful here even outside curriculum planning: decide what would count as evidence of the objective before building the instrument that scores it. Concretely — if a student met this objective, what would be sitting on the page in front of you?

Start from the objective’s verb

The fastest way in is the verb. Anderson and Krathwohl’s revision of Bloom’s taxonomy sorts learning objectives into six cognitive-process categories, and the verb is usually a direct signal of which one an objective is asking for:

• Remember — recall facts, terms, or sequences (list, name, define).
• Understand — grasp meaning (explain, summarize, classify, compare).
• Apply — use a procedure in a new situation (solve, execute, demonstrate).
• Analyze — break material into parts and see relationships (differentiate, organize, attribute).
• Evaluate — make and defend a judgment (critique, justify, check).
• Create — put elements together into something new (design, construct, generate).

Two problems show up in practice. Many objectives use “understand”-tier verbs that aren’t directly observable — “understand,” “know,” “appreciate.” You can’t watch a student understand something, only watch them do something that implies it; translate to an observable stand-in from the same tier. And the tier tells you what descriptors should be about: “remember” or “apply” objectives need descriptors about accuracy and completeness, while “analyze” and above need descriptors about the quality of reasoning or judgment, not just correctness — two students can reach the same correct answer while one shows far stronger analysis getting there. Naming the criterion after the translated verb (“Analysis of Causes,” “Evaluation of Sources”) keeps that cognitive level visible all the way through scoring.

Make it observable and measurable

A criterion only works if two people scoring the same work land in roughly the same place — which depends on whether descriptors point at something visible in the work, not something inferred about the student. “Shows understanding,” “tries hard,” and “demonstrates knowledge” describe a mental state you’re guessing at; “cites two pieces of textual evidence,” “shows each step of the calculation,” and “correctly identifies three of four variables” describe something on the page that any two graders can point to and agree on.

Brookhart’s guidance on rubric writing adds a related point: a criterion should describe the feature of the work that matters for the learning goal, not a proxy that merely correlates with it — don’t grade handwriting under an argumentation criterion, or let a lab write-up criterion drift into grading neatness instead of technique. Every descriptor should pass one test: could someone unfamiliar with the assignment point to the exact sentence or step that earns that score?

One criterion per objective

Keep criteria to one dimension each. A row labeled “Grammar and Creativity” forces a grader to average two unrelated qualities — a flawless but unoriginal paragraph and an inventive one full of run-ons land on the same score for entirely different reasons, and neither student learns anything specific from it. Two objectives usually mean two criteria, not one doing double duty.

The reverse mistake is over-fragmenting — turning every sub-skill into its own row until scoring one essay takes twenty minutes. Group only sub-skills that genuinely move together (“uses evidence” and “cites evidence correctly” can usually share one Evidence criterion). The target is enough criteria to cover what the objectives actually ask for, and no more.

Student-friendly language

Objectives and standards are usually written in curriculum language, aimed at teachers and administrators — “cite strong and thorough textual evidence to support analysis of what the text says explicitly.” That’s accurate and close to useless as rubric wording for a ninth grader checking their own draft before submitting it.

Translate it into language the student can read and act on: “Uses at least two direct quotes from the text and explains what each one shows.” Writing on how students use rubrics — including Heidi Goodrich Andrade’s work on rubrics and self-assessment — makes the same point: a rubric only does formative work if the student can read a descriptor and recognize whether their draft matches it. After drafting each descriptor, read it aloud and ask whether a student could use that sentence to check their own work; if not, it’s still standards language.

Worked examples across subjects

The translation holds the same shape across subjects — objective, verb, criterion, descriptor — even though the content looks completely different.

English

Objective: “Students will analyze how an author’s word choice shapes tone in a short story.” Verb: analyze. Criterion: Analysis of Word Choice and Tone. Strong: “Identifies at least two words or phrases and explains, with a quote, how each shapes the story’s tone.” Weaker: “Names a tone without connecting it to specific word choices.”

Science

Objective: “Students will design a controlled experiment to test a hypothesis.” Verb: design (create tier). Criterion: Experimental Design. Strong: “Identifies the independent and dependent variables, names a controlled variable, and the procedure could be reproduced by another student.” Weaker: “A procedure is present but doesn’t identify variables or control for other factors.”

Math

Objective: “Students will solve multi-step equations and justify each step.” Verb: solve, with a secondary justify. Criterion: Solution and Justification. Strong: “Every step is shown in order, and each names the operation or property used.” Weaker: “The final answer is correct but intermediate steps are missing or unexplained.”

In all three, the descriptor never restates the objective — it says exactly what has to be visible in the submission for that level to apply.

Aligning descriptors to levels

Once you have a criterion, its descriptors need to describe the same dimension at different strengths across levels — not different dimensions. A common failure: the top level talks about originality, the bottom about neatness. Those are two different criteria wearing one label, and a grader has no consistent way to apply them.

Brookhart’s advice on parallel descriptors is the fix: hold the sentence structure constant across levels and change only the degree or quality words. For the English example above, “identifies two word choices and explains each with textual evidence” (strong) against “identifies one word choice with a general explanation” (developing) against “names a tone without identifying word choices” (beginning) are parallel — same criterion, same structure, different degree.

It also matters which tier the lowest level lands on. If the objective calls for analysis, even the weakest descriptor should still describe weak analysis, not drop to an unrelated skill like spelling — otherwise a student at the bottom has no idea what to do differently next time. If these levels are meant to track proficiency over time against a standard, rather than score one assignment, see Standards-Based Grading With Rubrics for how proficiency scales are typically built and reported.

Build it

The process, start to finish: take the objective, find its verb, place it on Bloom’s six-tier scale, name a criterion after it, write one descriptor per level in parallel language, then translate every descriptor into wording a student could read and use. Pilot it on two or three real submissions before trusting it for a whole class set — real student work surfaces a vague descriptor faster than reading it in the abstract does.

Start from our free Blank Rubric Template if you’re converting objectives from scratch, or the Essay Rubric if the objective in front of you is about reading or writing — both are ready to edit with your own criteria and export as soon as the descriptors are done.