How to assess a learning project with a clear rubric
A learning project rubric connects the task you intended to complete with observable evidence. Start with what the user should be able to do, write a small set of checks and assess the work against them. This guide is for beginners who can finish an exercise but struggle to explain whether they have learned something transferable.
A learning project rubric connects the task you intended to complete with observable evidence. Start with what the user should be able to do, write a small set of checks and assess the work against them. This guide is for beginners who can finish an exercise but struggle to explain whether they have learned something transferable.
Key ideas
- The task names a user and an observable end state.
- Correctness and explanation have separate criteria.
- Each judgement points to an actual test record.
- An unchecked criterion remains visibly unchecked.
Define the task before the grade
Choose a result a second person can attempt, such as finding a lesson and opening its practice file. ‘Make a beautiful website’ leaves the reviewer guessing. ‘Find the Python exercise from the course page without help’ gives the project a visible purpose. Add the starting point, required information and expected end state. A rubric should describe this particular task, rather than every possible skill in the field.
Separate correctness, explanation and usability
An exercise can run correctly while the learner cannot explain a decision. It can also contain a thoughtful explanation yet fail for someone using a keyboard. Give these questions separate rows. Use plain levels such as missing evidence, partly demonstrated and demonstrated on the specified task. Avoid turning a private learning worksheet into a professional qualification or an accessibility certification. W3C Easy Checks is explicitly a starting review, not proof of complete accessibility.
Attach evidence to every judgement
A screenshot shows appearance; a short test record shows a particular behaviour. Save the starting URL, action, expected result and observed result. Include a failed attempt and its correction when that explains your reasoning. MDN’s testing material supports testing real user situations; the rubric here is our editorial learning method. Do not label an unchecked item ‘passed’ just because the final screen looks convincing.
Use the rubric to choose the next exercise
Ask a peer to follow the task without your narration. Record where your instructions or interface require guessing. Then choose one weakness to improve and repeat the same check. Keep earlier evidence so you can compare your decisions, not merely two polished screenshots. A low result is useful when it points to a specific missing skill. A high self-assessment still describes this exercise and cannot predict a job offer or client outcome.
Your evidence checklist
Mark only what you have checked. This records your own progress, not an independent audit or a predicted result. There is no automatic saving; download the note if you want to keep it.
0 / 6 checked
In everyday language
Assess the work against its stated task, keep the evidence and choose one improvement you can test again.
Try it yourself
Create a rubric for a two-page course website. Use three rows: finding the exercise, opening the file and explaining the navigation choice. Test each row once, attach the evidence and revise the row that needs guessing.
Expected result
A small table with three criteria, clear evidence links and one justified next action. The example is fictional and the assessment covers this exercise only.
Check your answer: Does a self-assessment prove professional competence?
It shows how you judged one defined exercise. Professional work adds unfamiliar requirements, collaborators and operational constraints. Keep the rubric and test records as evidence of your reasoning, then try another task before making a broader claim about competence.
Questions
Does a self-assessment prove professional competence?
It shows how you judged one defined exercise. Professional work adds unfamiliar requirements, collaborators and operational constraints. Keep the rubric and test records as evidence of your reasoning, then try another task before making a broader claim about competence.
Should every row receive a numeric score?
Numbers help only when the levels have clear meanings. Missing, partly demonstrated and demonstrated can be enough for a beginner. The important part is the observation behind the level and the next action it suggests, rather than the appearance of precise measurement.
Assess the work against its stated task, keep the evidence and choose one improvement you can test again.
Sources and further reading
- MDN — Testing strategies ↗Sources checked:
- GOV.UK — In-depth research interviews ↗Sources checked: