Configuring peer reviews¶
Peer reviews are a powerful learning method: learners assess each other's work and learn twice in the process — while writing their own answer and while attentively reading others' answers. This page shows how to set up a peer-review task in a course.
When peer review makes sense¶
Peer review fits well for:
- Reflection tasks — one's own experiences, personal application of the learning material
- Practical examples — learners should submit their own example or project
- Soft skills (communication, argumentation) — where there's no "right/wrong"
- Larger classes — where individual trainer grading doesn't scale
For pure knowledge tests, classic multiple choice or AI free-text grading is more efficient.
Creating a peer-review task¶
Peer review is configured as its own lesson type within a course.
- Create a new lesson in your course.
- Choose lesson type Peer-review task.
- Write the task — clearly formulated, stating what learners should submit.
- Choose the answer format:
- free text (typed in the browser)
- file upload (PDF, image, ZIP, …)
- both
- Set the number of reviewers per submission — typically 2-3.
- Define the grading criteria (see the next section).
- Choose the anonymization — default: reviewers and submitters don't know each other.
- Set the deadline — when submissions + reviews must be done.
- Save.
Screenshot to follow
Peer-review configuration with task and criteria
Defining grading criteria¶
Peer-review assessments are not a simple "1-5 stars". You define a set of criteria per task. Three question types can be combined:
Scale questions¶
A question with a scale of 1-5 (or 1-10).
Examples: - "How clear is the argumentation?" 1 (very unclear) — 5 (very clear) - "How well was the material applied?" 1-5 - "How comprehensible is the reflection?" 1-5
Yes/no questions¶
A binary checklist — either met or not.
Examples: - "Were all requirements from the task addressed?" - "Are sources cited correctly?" - "Is the submission formally complete?"
Free-text feedback¶
Reviewers give qualitative feedback in a free text field. For the submitting person, this is often the most valuable.
Recommendation: mix the question types — 2-3 scale questions for structured grading, 1-2 yes/no questions for formal requirements, and one mandatory free-text field for qualitative pointers.
Anonymization¶
Three levels:
| Mode | Who sees what |
|---|---|
| Fully anonymous (default) | The reviewer doesn't know the submitter, the submitter doesn't know the reviewers |
| Reviewers anonymous | Submitters see the assessment content, but not from whom |
| Open | All names visible — encourages direct exchange, but can produce self-censorship |
Recommended in most cases: fully anonymous. Learners give more honest feedback when they're not personally connected.
Reviewer assignment¶
- The system assigns reviewers automatically as soon as a person submits a piece of work.
- Reviewers come from the pool of all enrolled learners — the submitting person is excluded.
- Per submission, the number of reviewers set in the setup.
You don't need to do any manual assignment as a trainer. In very small classes (< 4 learners) the pool isn't enough — then the reviewer assignment turns out sparser, or individual reviewers get several pieces of work.
Workflow for learners¶
Two phases run per learner:
- Submit: upload or type your own answer, mind the deadline
- Assess: read the assigned work of others and rate it by the criteria
Learners see the assessment tasks under Peer reviews in the main navigation. Detailed user guide: for learners: peer reviews.
Reviewing the reviews — as a trainer¶
As a trainer you see:
- which submissions were made when and by whom
- which reviews were written when and by whom
- where reviews are missing (e.g. because a reviewer missed the deadline)
In Analytics → Lesson breakdown → Peer reviews you'll find all data on the task.
When reviews are unusable¶
Sometimes weak feedback comes from reviewers ("all good", or worse: "dumb answer"). You can:
- Mark individual assessments as invalid — they don't count toward the average
- For offensive content: delete the assessment and, if needed, talk to the reviewer
- For systematically weak feedback: build in a workshop idea "how to write good peer-review feedback" as the next course iteration
Frequently asked questions¶
What happens if a reviewer misses the deadline? The assessment stays open; the reviewer can still submit it after the deadline. Trainer analytics shows who is overdue. For repeated misses: follow up organizationally or swap out the reviewer.
Can I assign reviewers manually? Currently not supported in the UI — assignment is automatic. If a particular pairing doesn't work (e.g. because two people don't get along outside the academic context), you can offer the submitting person an additional trainer assessment as a "human double-check".
How many reviewers per submission make sense? - Two is often enough, gives two perspectives - Three smooths out outliers (a very strict or very lenient assessment is balanced out) - Four or more is hardly worth it, the effort per learner gets too large
Do I see which reviews received which ratings — as an assessment of the reviewers? There's no direct "reviews of the reviews". If good review practice matters to you, give reviewers some spot-check feedback yourself — e.g. send a direct message "Your feedback to person X was especially constructive".
Learners feel unfairly assessed — what should I do? First step: review the assessment yourself. If the reviewer assessment is really off, add your own trainer assessment as a "corrective". For repeated fairness discussions: perhaps your criteria set isn't clear enough — adjust the grading criteria.