Why peer review gets harder as the class grows

In a small class, quality control is informal. The instructor reads a few reviews, notices that one student has written a single line, and has a quiet word. None of that scales. In a large course, three problems appear at once.

The first is logistics. Allocating reviewers by hand, chasing late submissions and making sure nobody ends up reviewing an empty document becomes a full time job. The second is variability. With four hundred students you will have brilliant reviewers and students who write "good job, well structured" and move on. Without a mechanism to raise the floor, the weakest reviews define the student experience, because a student who receives two useless reviews concludes that the whole activity is a waste of time. The third is trust. Large courses are usually where grades feel most impersonal, so students are quickest to suspect that peer review is a way for staff to avoid marking.

Everything below is aimed at those three problems.

Decide what the review is for before you design it

The most common mistake is to run peer review as a substitute for instructor marking. That framing invites every objection students have, and it is also the least interesting use of the activity.

Research on peer assessment consistently finds that the biggest gains come from giving feedback rather than receiving it, because producing a judgement against criteria forces students to internalise what quality looks like. If that is the goal, then the review itself is the learning activity, and it should be treated like one. It gets time in class, it gets modelled, and it gets credited.

So write down one sentence before you build anything. Either the aim is formative improvement of a draft before a final submission, or the aim is teaching students to evaluate work against professional standards, or the aim is generating a defensible individual mark. Those three aims produce very different designs, and trying to serve all three at once is how peer review ends up serving none of them.

Build criteria students can actually apply

Criteria written for staff do not work for students. "Demonstrates critical engagement with the literature" means something to a marker with ten years of experience and almost nothing to a second year student at eleven at night.

Rewrite each criterion as an observable question. Instead of asking about critical engagement, ask whether the author explains why each source was chosen and whether they point out any limitation or disagreement between sources. Instead of asking about structure, ask whether the reader can state the main argument after reading only the introduction and conclusion.

Keep the list short. Four to six criteria is usually the ceiling for an undergraduate reviewer. Every criterion beyond that reduces the care given to all the others. Pair each rating scale with a required comment box, because a number on its own tells the author nothing about what to change. If you are building criteria from scratch, the Rubric Assistant can draft a first version from your learning objectives, which you then edit rather than write cold.

Choose your allocation settings deliberately

Three settings matter more than the rest.

Number of reviewers per submission. Two is the minimum that lets an author see a pattern rather than one person's opinion. Three is better and is the point at which an outlier review stops being distressing, because the author can see it is an outlier. Four starts to cost more student time than it returns.

Anonymity. Anonymous review usually produces more honest criticism, which is what you want when the work is a draft. Named review usually produces more careful, professional phrasing, which is what you want when the point is to build a reviewing culture in a cohort that will work together for three years. Pick based on the aim you wrote down, and tell students which one you chose and why.

Timing. Peer review only improves work if there is time to act on the feedback. A review window that closes the day before the final deadline is decoration. Leave at least a week between the close of reviewing and the final submission, and build the revision into the assessment by asking students to note what they changed.

Raise the floor on review quality

This is the part most courses skip, and it is the part that decides whether students believe in the activity.

Start with a calibration exercise. Before students review each other, give the whole class the same anonymised sample of work, ask them to apply the rubric, and then show them how you would have applied it. Ten minutes of this in a lecture does more for review quality than any amount of written guidance, because students discover for themselves that they were grading generously or missing a criterion entirely.

Second, give students a shape to follow. A simple pattern such as naming one thing that works, one thing that would most improve the piece, and one concrete suggestion for how to do it, prevents the empty praise that makes students cynical. The Feedback Coach does the same job at the moment the student is typing, which is when the advice lands.

Third, assess the reviewing. If review quality carries no credit, a rational student under pressure will write the shortest review that passes. A small weighting, five or ten percent of the assignment, is enough to change behaviour without turning the activity into another high stakes assessment.

Plan the exceptions before the week begins

At scale, exceptions are not rare events, they are a predictable percentage of the cohort. Decide in advance what happens when a student submits nothing, when a submission is unreadable, when a review is abusive, and when a student misses the review window because of illness.

The useful habit is to write these rules into the assignment description rather than handling them case by case in email. Automatic reallocation when a submission is missing, a clear route for a student to flag an inappropriate review, and a stated policy on missed reviews will save your teaching team more hours than any other single decision. If you want to know more about how reallocation, reminders and missing submissions are handled in the tool, our Peer Review FAQ walks through the common cases.

Use the data you already have

A peer review cycle generates a picture of the cohort that is hard to get any other way. Before the next class, look at which criterion produced the widest spread of ratings, because that is the criterion students do not understand yet, and it is worth ten minutes of teaching. Look at the submissions that received unusually low ratings from every reviewer, because those are the students who need an early conversation. Look at the reviewers whose ratings sit far from everyone else's, because that is either a student who has misread the rubric or a student who has spotted something the others missed.

Exportable analytics make this a fifteen minute task rather than a research project, and it turns peer review from an activity you run into a source of evidence about how the course is going.

What this looks like when it goes well

The pattern that repeats across institutions doing this well is not complicated. The criteria are short and concrete, students practise on a sample before they review each other, there is real time between feedback and revision, the reviewing itself carries a small amount of credit, and the exception rules are written down before the semester starts.

None of that requires a new pedagogy. It requires making explicit the things a good instructor does instinctively in a room of twenty, so that they still happen in a cohort of four hundred.

Where to go next

If you would like to see how this works inside the tool itself, the Peer Review solution page walks through rubric building, reviewer allocation, anonymity settings and the automated reminders that take the chasing off your hands. The Peer Review FAQ is a good next stop if you have the practical questions that usually come up before a first run, such as how the activity sits in your LMS, what students see, and what happens when someone misses a deadline.

It is also worth looking at the tools that sit naturally alongside this one, including Self Assessment, Assignment Review and Skill Review. You can see how they all fit together on the FeedbackFruits tool suite page, or read more about what a feedback and assessment solution actually is if you are building the case for your institution.

And if you want to keep reading on this topic, we have Peer assessment: methods and examples, How to create effective peer feedback criteria and Why feedback and assessment is so hard to do well at scale.

‍