Assessment & feedback · 5 min read
Peer Assessment That Actually Works
The common mistakes and the practices that produce genuine peer learning
Published 2026-05-27
Peer assessment — pupils evaluating each other's work — has a strong evidence base when done well and a weak one when done poorly. The research consistently shows large benefits from well-implemented peer assessment and near-zero benefits from the typical primary school version.
The typical primary school version: one pupil writes something, another pupil reads it and writes 'two stars and a wish', the paper is returned, nothing changes. This is not peer assessment — it is peer commenting. The distinction matters.
Why most peer assessment fails
Pupils haven't been taught what good looks like. Before pupils can evaluate work, they need to have a clear, internalised understanding of what quality looks like in this task. 'Give two positives and an improvement' is not a quality standard — it is an empty scaffold. A pupil who doesn't know what good writing looks like cannot assess whether someone else's writing is good.
The feedback is too vague to act on. 'Good detail' is not actionable feedback. 'The second paragraph has three specific examples, but the first one doesn't have any — could you add one?' is actionable. The specificity required for useful peer feedback is the same as for useful teacher feedback.
There is no response phase. Peer assessment without a response task — where the author acts on the peer feedback — is a commentary exercise, not an assessment exercise. The learning happens when the author revises in response to feedback.
What works
Success criteria as the assessment tool. When pupils have clear, specific success criteria before writing, peer assessment becomes straightforward: 'Does this piece meet each criterion?' This is concrete enough for primary pupils to do meaningfully.
Peer assessment with modelled examples first. Showing the class a piece of work and modelling peer assessment of it — aloud, together — teaches the skill before applying it independently.
Focused feedback: one thing. Rather than comprehensive feedback, ask peers to identify one specific thing. 'Find one sentence in this piece that you think could be improved, and explain why.' This is more achievable and more useful than an open-ended evaluation.
Built-in response time. After receiving peer feedback, pupils have five minutes to make one specific change to their work. The change makes the feedback real.
Free bundle for this topic
KS2 Maths Pack
10 maths resources designed for retrieval practice and low-stakes review.
Keep reading
Assessment & feedback
Formative Assessment Without the Workload
Formative assessment — finding out what pupils understand in real time — does not require marking. The most effective techniques take less than five minutes and produce more useful data than a marked worksheet.
6 min read
Assessment & feedback
The Problem with Marking Grids — and What to Do Instead
Marking grids — structured rubrics listing criteria, usually with RAG ratings — have become ubiquitous in primary marking. They also frequently produce worse feedback than a well-written sentence. Here's why, and what works better.
5 min read