Once you have collected student work samples, you need a consistent way to evaluate them. A rubric gives everyone on your faculty team a shared standard so that the same piece of work gets scored the same way regardless of who is reviewing it.
What a Rubric Does
A well-designed rubric defines levels of achievement for each outcome dimension, making it possible to compare student performance across courses, sections, and years. It removes subjectivity from the scoring process and makes your findings defensible.
Example Rubric Structure
| Outcome dimension | 1 — Beginning | 2 — Developing | 3 — Proficient | 4 — Exemplary |
|---|---|---|---|---|
| Critical thinking | Identifies a problem but does not analyze it | Analyzes problem with limited evidence | Analyzes problem using relevant evidence | Synthesizes evidence to form original conclusions |
| Written communication | Ideas unclear; structure difficult to follow | Ideas present but organization inconsistent | Ideas clear and logically organized | Exceptionally clear; engaging and precise |
| Use of sources | Few or no sources; citations missing | Sources used but not well integrated | Sources well chosen and properly cited | Sources critically evaluated and expertly woven in |
Each row is one dimension of the outcome. Each column is a performance level. Faculty score each dimension independently, then look at patterns across students.
How to build your rubric
1. Identify the dimensions of your outcome.
Break the outcome into 2–5 observable components. For example, “written communication” might include clarity, organization, use of evidence, and mechanics. Each component becomes a row in your rubric.
2. Describe what each performance level looks like.
Write concrete, behavioral descriptions for each level (typically 3–4 levels). Focus on what is observable in the student work — not effort, attitude, or potential. Avoid vague language like “adequate” or “good.”
Start by describing “Proficient” — what does meeting the standard look like? Then describe what falls short and what exceeds it. This is easier than starting at the extremes.
3. Pilot with a small sample.
Have two or three faculty independently score the same 3–5 pieces of student work using your rubric draft. Compare scores and discuss disagreements. Revise the language anywhere raters consistently disagree.
4. Set your expected level of achievement.
Before scoring your full sample, decide as a faculty team: what percentage of students should score “Proficient” or above? This benchmark is what you will compare your actual results against when you interpret findings.
5. Score your sample and look for patterns.
Score student work samples using the finalized rubric. Tally results by dimension and performance level. Look for where students cluster — which dimensions show strength, and which show consistent gaps across the group.
Rubric options — build your own or start from an existing one
VALUE rubrics
Pre-validated rubrics from the Association of Colleges and Universities (AAC&U) cover 16 learning areas including critical thinking, written communication, quantitative literacy, and teamwork. These rubrics are strong starting points for the development of common outcomes.
URI-developed rubrics
These examples were built by URI faculty for specific programs including information literacy, thesis defense, and graduate communication. They are useful models when building rubrics from scratch.
Common Questions
Does the rubric need to be created specifically for assessment?
Not necessarily. If you already use a grading rubric for an assignment, you can adapt it for program assessment — as long as it scores the outcome dimensions you care about at the program level, not just course-level grading criteria.
Do all faculty have to use the same rubric?
Yes. A shared rubric is what makes it possible to compare student work across sections and draw conclusions at the program level.
How many students do we need to score?
Ideally all students completing the key assignment. For larger programs, a random sample of around 20% is generally sufficient, as long as it is representative. The goal is to collect enough data to identify reliable patterns, not to evaluate every individual.
What do you want to do next?
← Back: SLOAA Homepage
← Back: Program Assessment