The prompt
Score 10 of 10, Excellent
Use it when You set an assignment or task and need a rubric that makes grading consistent and tells learners what good work looks like before they start.
- {{assignment}}
- {{learners}}
- {{objectives}}
- {{scale}}
Full prompt and why it works
Prompt
<assignment> {{assignment}} </assignment> <context> Learners: {{learners}} What the assignment is meant to show: {{objectives}} Scoring scale: {{scale}} </context> <constraints> - 4 to 6 criteria, each assessing something different and taken from the objectives. Do not include criteria the assignment does not ask for, such as neatness or length, unless the objectives name them. - For each criterion and level, describe what the work does in observable terms, for example "Cites at least two sources and connects each one to the argument", not "Good use of sources". - Levels differ in the quality of the work. Quantity words alone, such as "some" and "many", do not separate one level from the next. - Weight the criteria to match the objectives, as percentages that add up to 100. - Write the descriptors so that a learner could use them to check their own work before handing it in. </constraints> <format> 1. A Markdown table: one row per criterion, one column per level, plus a Weight column. 2. A learner-facing checklist of 5 to 8 lines, written as "I have..." statements. 3. Grading notes: two or three borderline cases a grader is likely to meet, and how to score them. </format> <task> Based on the assignment and context above, write a rubric that two different graders would apply to the same piece of work and reach the same score. If the objectives or the scale are unclear, assume the most likely reading and state your assumption in one line above the table. </task>
Variables
- {{assignment}}
- The assignment brief as learners receive it.
- {{learners}}
- Who does the assignment, for example "10th grade history students" or "new analysts in their first month".
- {{objectives}}
- What the work should demonstrate, for example "build an argument from primary sources".
- {{scale}}
- For example four levels from Beginning to Exceeds, or points out of 20.
Why it works
- What good looks like. The bar is that two graders reach the same score, which forces descriptors that name observable features instead of words like "good" and "some".
- An example. One example contrasts an observable descriptor with a vague one, which shows the level of precision wanted in every cell.
- Who it's for. Descriptors a learner can check their own work against make the rubric useful to learners before they hand work in, as well as to graders afterward.