When Assessment Starts to Teach

Student self-assessment and peer-assessment are often introduced as ways to make learners more active. The more consequential question is narrower: when does a student’s judgment become usable feedback rather than another score? This series examines self-assessment and peer-assessment interventions through that question, so that a reader can tell the difference between an assessment that ends with a verdict and one that gives a learner a defensible next move.

That distinction matters because the same label can describe very different classroom experiences. A student may circle a number predicting a grade, write comments against a rubric, review a classmate’s argument, or compare a draft with a model before revising it. All of these can be called assessment. They do not all give the learner the same information, create the same incentives, or support the same kind of learning.

The problem with a score that ends the conversation

Imagine two students receiving the same writing task. The first reads the teacher’s score, records it, and moves on. The second compares the draft with a set of criteria, marks where the claim is unclear, asks whether the evidence supports it, reads a peer’s comment, and revises before submitting the next version. Both students have been assessed. Only the second assessment necessarily creates information that can change the work.

This is the basic tension in student-directed assessment. A grade or rating is a judgment, but a judgment is not automatically feedback. Feedback has a practical direction: it tells someone something about a process or product that can inform an adjustment. If the learner has no opportunity to correct, practise, rethink, or choose a better next task, the educational purpose of the judgment is weakened. The assessment may still document performance, but it has not yet become a learning process.

The distinction is especially important for self-assessment. In a critical review of research, formative self-assessment is defined by its purpose of generating feedback that improves learning or performance. Summative self-assessment, by contrast, may ask students to judge their competence or assign themselves a final mark after the work is finished. Neither form is imaginary, and summative self-assessment is not automatically worthless. But the incentive changes when the student’s rating can affect the final grade. Research summarized in the review found that self-ratings became more inflated in that situation, particularly among weaker performers. The judgment was being used to negotiate an outcome, not primarily to identify a next learning action.

That finding does not prove that students are incapable of judging their own work. It shows why purpose and context belong in the definition. A learner who is asked, “What grade do you deserve?” is answering a different question from one who is asked, “Which criterion is not yet met, and what will you change?” Treating the answers as the same kind of self-assessment produces a misleading verdict about the method.

Peer-assessment has the same problem in a different form. Asking students to assign a classmate a score can make assessment visible, but the score alone does not explain what the reviewer noticed or what the author should do. A peer comment tied to a criterion can provide a second perspective; a peer grade detached from reasons may simply add another judgment to the pile. The central issue is not whether the evaluator is the teacher, the learner, or a classmate. It is whether the evaluation produces evidence that someone can interpret and use.

flowchart LR N01["Learning purpose and criteria"] --> N02["Student judges own or peer work"] N02 --> N03["Feedback and interpretation"] N03 --> N04["Revision or strategy change"] N04 --> N05["Next attempt and new evidence"] N05 --> N01 N02 --> N06["Terminal score"] N06 --> N07["No required next action"]

The diagram captures the difference in one glance. A terminal score can be accurate and still leave the learning process unchanged. A formative judgment can be imperfect and still prompt a productive revision, provided the learner has a meaningful standard, interpretable evidence, and a chance to act.

What self- and peer-assessment actually mean

Self-assessment is not one narrow technique. It can concern a student’s competence, the process used to learn, or a product such as an essay, presentation, solution, or performance. It can involve a quick indication of understanding, a self-test, a reflective journal, a rubric-referenced critique, or a self-grade. The useful question is not whether the student supplied a number. It is what the student was asked to notice, what standard made the notice meaningful, and what decision followed.

Peer-assessment is similarly broad. Students may rate one another’s work, write comments, discuss a solution, or use shared criteria to explain why a product does or does not meet a standard. The distinctive feature is that students take both assessor and assessed roles. Reviewing someone else’s work can require the reviewer to apply criteria and justify a judgment; receiving a peer’s perspective can expose the author to evidence they did not generate alone. The two roles can teach different things, even when they occur in the same activity.

The practices nevertheless share a common structure. The learner needs some sense of the goal, a way to recognize relevant evidence, and a route from the judgment to an adjustment. In self-assessment, the evidence may be an internal observation of the learner’s process or a comparison between their work and a standard. In peer-assessment, it may include another person’s interpretation of the same work. In both cases, the assessment is potentially formative because it helps the learner monitor progress, identify a gap, and direct future effort.

Criteria are what make that structure more than a vague invitation to reflect. A criterion such as “make the work better” gives little guidance. A task-specific criterion such as “state the central claim clearly and support it with relevant evidence” gives the learner something to locate in the draft. Concrete and graduated standards can also make small improvements visible. They do not eliminate judgment; they give judgment an object that can be examined and discussed.

This is why the research literature often treats rubrics, checklists, models, and scripts as part of the learning environment rather than as administrative accessories. In examples summarized by the critical review, students inspected drafts against criteria, identified evidence that a standard had or had not been met, and recorded changes for a later version. The important event was not the existence of a rubric. It was the learner’s use of the rubric to connect an observation to an action.

The distinction also protects against an unhelpful form of self-assessment: a global statement about the self. “I am bad at mathematics” may express discouragement, but it does not identify which concept, operation, or strategy should be addressed. A task-specific observation—“I can set up the equation, but I lose the sign when simplifying”—has more leverage because it narrows the next decision. Formative assessment is strongest when it describes work and process in a way that leaves room for change, rather than converting a temporary performance into a fixed identity.

Peer-assessment adds a social condition to this shared structure. A peer’s comment can be easier to understand because it comes from someone working on a similar task; it can also be difficult to accept if the relationship feels unsafe or the criteria are unclear. Trust, privacy, accountability, and reviewer preparation are therefore not peripheral details. They affect whether external feedback becomes evidence the learner can interpret. A peer process that produces comments without a credible way to use them may be socially busy but educationally thin.

Why the subject matters beyond immediate performance

The case for self- and peer-assessment is broader than the claim that they raise a test score. The recent meta-analytic literature separates outcomes into self-regulated learning strategies—such as metacognitive strategy use, effort regulation, and management of resources—and affective outcomes such as attitude, emotion, motivation, and self-efficacy. These are not interchangeable measures. A student who becomes better at monitoring a task is showing a different change from a student who feels more capable or persists longer.

A 2026 synthesis of 75 studies and 353 effect sizes reported positive average effects for both self-assessment and peer-assessment across these broader outcome groups. Against control conditions, self-assessment had an average effect of 0.293 for self-regulated learning outcomes and 0.393 for affective outcomes; the corresponding peer-assessment estimates were 0.468 and 0.325. Those figures make an important point: the practices can matter beyond grades. They do not make a second point that would be equally important but false—that every activity carrying either label produces those results.

The same review notes that some outcomes were measured only once, that evidence was sparse for several categories, and that the connection between short-term changes in strategy or feeling and sustained long-term learning remains a theoretical prediction. A separate experimental meta-analysis focused on motivation also found positive average effects for self-assessment and peer-assessment, with no significant difference between them. Yet its large between-study variation shows why an average should be read as a signal across many arrangements, not as a property of an abstract method.

This evidence changes the practical question educators should ask. Instead of “Should I use self-assessment or peer-assessment?” the more useful questions are: What learning process should the student be able to regulate? What evidence will make that process visible? Which form of judgment gives the learner a perspective they can use? What must happen after the judgment? And what social or instructional support will make that next step possible?

Those questions keep the subject in view while avoiding two opposite mistakes. One mistake is to treat self- and peer-assessment as magic labels: add a rating box or a peer-review round and assume learning will follow. The other is to dismiss them because students sometimes misjudge their performance. A critical review argues that classroom research should distinguish consistency with an external rating from the educational usefulness of the judgment. A learner’s alignment with a teacher’s score is worth studying, but it is not the whole purpose of formative assessment. The harder and more consequential question is whether the judgment improves the learner’s next choice, revision, or strategy.

That is the threshold at which assessment starts to teach. A score can close a record; a learning process opens a decision. The next task is to examine the links between them: how goals and criteria shape a judgment, how feedback is interpreted, and how a learner turns evidence into a change in work or behaviour. Without those links, self-assessment and peer-assessment remain names for activities. With them, they become ways for students to participate in the regulation of their own learning.

References