UDL and learning diversityAssessment and academic integrityLesson and material designEquity, bias and the digital divide

Design a differentiated assessment using UDL principles

Tested on
Claude Sonnet 4.6, November 2026
Estimated time
20 min
Time saved
2-3 hours
Published
2026-11-09
Last reviewed
2026-11-09
Attribution
Equipo Circles

Context

A primary or secondary teacher who has an assessment instrument — written test, practical assignment, oral presentation — and needs to adapt it for students with special educational needs (SEN) without creating a completely separate version or reducing the cognitive demand of the learning objective, before the assessment with a draft already in hand.

Originally written for Chilean classrooms. References to PIE (Special Education Integration Program), Bases Curriculares, and Law 19.628 apply specifically to the Chilean system. Adapt to your local inclusion framework, special education program, and data-protection regulations.

Paste first

Before opening the model, have ready:

  • The learning objective or assessment indicator you want to measure, as it appears in your national curriculum or your unit plan
  • The current assessment instrument or draft you have (copy it as-is, without any student data)
  • A functional description of the SEN group: what barriers do they have to demonstrating their learning? (slow reading pace, difficulty with extended writing, speech-language needs, autism, other)
  • The type of assessment: is it formative or summative? Does it count toward a grade?
  • Practical constraints: time available, materials, whether a support teacher or special educator will be present that day

Privacy rule: Do not paste names, government IDs, individual clinical diagnoses, or any confidential special-education records such as Individual Support Plans or medical certificates. Describe the group's characteristics functionally: "two students take more than twice as long to read the questions and struggle to structure paragraphs," "one student understands the content well but cannot express ideas in writing coherently." Your school's data-protection regulations cover health and diagnostic data — do not enter that information into any LLM.

Prompt

Act as a Universal Design for Learning (UDL) specialist with experience in real primary and secondary classrooms. I need to design a differentiated assessment that measures the same learning objective for the whole class, but offers different ways to demonstrate that learning — especially for students with special educational needs — without reducing the cognitive demand.

**Subject and grade level:** {{subject}} — {{grade level, e.g.: Year 5, Grade 10}}

**Learning objective I want to assess:**
{{paste the exact learning objective from your national curriculum or unit plan}}

**Current instrument (or draft):**
---
{{paste the full assessment here, without any student data}}
---

**Functional description of the SEN group (no names or individual diagnoses):**
{{describe the barriers to demonstrating learning functionally — e.g.: "two students with slow reading pace and difficulty with extended writing, one student with speech-language needs who understands the content but cannot express ideas in writing coherently, the rest of the class with no reported SEN"}}

**Type of assessment:** {{formative / summative — indicate whether it counts toward a grade and what percentage}}

**Practical constraints:** {{time available, materials, whether a support teacher or special educator will be in the room that day}}

**What I need:**

1. A brief analysis of what concrete barriers the current instrument presents for the described SEN group (maximum 3 barriers, ranked from highest to lowest impact).
2. For each barrier, one or two expression alternatives that maintain the same learning objective and the same level of demand, changing only the format of demonstration.
3. A rewritten version of the full instrument that integrates the alternatives — without duplicating the assessment, but offering options within the same instrument so that everyone works from the same page.
4. A brief note on how to present the options to the class without it appearing that some students have an "easier" version.
5. A single rubric or achievement criterion that can be used to assess all variants of the instrument against the same standard of achievement.

Do not simplify the conceptual content or reduce the indicators that measure the learning objective. Differentiation is in the channel of expression, not in the demand.

Expected output

Example (abbreviated):

Differentiated assessment with UDL — Science Learning Objective | Year 5. Both versions are scored out of 15 and graded with the same rubric.

Version A — standard (28 students): Question 1 (3 pts): function of the digestive system + 2 organs + explanation of each. Question 4 (5 pts): choose a different system; explain (a) main function, (b) relationship with another system and the mechanism of connection, (c) consequences if that system fails.

Version B — adapted for SEN (4 students): The same questions with scaffolding: labeling diagrams with a sentence frame ("The _____ is used to _____"), a word bank for the nervous and musculoskeletal systems, a pre-structured diagram with arrows to explain the circulatory-respiratory relationship.

Unified rubric: assesses demonstrated concepts, not form of expression. An accurately annotated diagram is equivalent to an accurately written response; there is no different score by version.

Protocol: the teacher draws the diagram for Question 2 on the board before beginning; between minutes 35–43, asks a brief oral clarifying question to the SEN students and records the answer in the margin.

Watch out for

  • UDL does not authorize lowering the curriculum standard. If the model suggests removing questions, reducing the conceptual level, or removing learning objective indicators for SEN students, reject the suggestion and ask it to maintain the demand while only adapting the format for access or expression.
  • Do not paste Individual Support Plans, diagnostic reports, or any confidential special-education document. Those records are protected by your school's data-protection regulations and must not be entered into any LLM. Describe only what you observe in the classroom: what the student can do, what they find difficult, what supports work.
  • The model may suggest alternatives that require technology you do not have (audio recording on a tablet, text-to-speech software, specific apps). If so, ask for "low-tech alternatives, on paper or with basic classroom resources, that I can apply on my own without additional equipment."

Suggested iteration

If the proposed alternatives are too similar to each other and do not resolve the main barrier, ask: "The oral option and the written option are too alike. Give me a third alternative in a completely different format — for example, a visual diagram, an image sequence with captions, or a practical demonstration — that still measures exactly the same learning objective." If the proposed rubric favors the written version over the others, ask: "Rewrite the 'analysis' criterion in the rubric so that it is format-neutral: so it applies equally to someone who responds in writing, orally, or with a graphic organizer."