The Judgment-to-Learning Loop

The subject here is the mechanism inside student self-assessment and peer-assessment interventions: how a learner moves from a learning goal to evidence, judgment, and a changed next move. The focus matters because an assessment can look active while leaving learning untouched. To judge whether self- or peer-assessment is likely to teach, trace the chain and locate the point at which information becomes action.

A useful starting distinction is between a result and a reason. “You earned 72” is a result. “Your claim is clear, but two paragraphs introduce evidence without explaining how it supports the claim” is a reason. The second statement gives the learner something to inspect and change. Even then, it becomes formative feedback only if the learner has an opportunity and a motive to use it.

Criteria turn impressions into inspectable evidence

Learning goals describe what a student is trying to achieve. Criteria make the goal visible in the work. They name features that can be looked for, discussed, and improved. Without that reference point, self-assessment easily becomes a feeling—“I think this is good”—and peer-assessment becomes a preference—“I like this version better.” Neither response necessarily tells the learner what the task requires.

Consider a student writing an explanation of a scientific process. “Write a strong explanation” is a goal, but it is too broad to guide an inspection. Criteria might ask whether the explanation states the central process, orders its stages accurately, uses evidence for the important claim, and makes the relationship between cause and effect explicit. The student can now point to a sentence, compare it with the criterion, and identify a gap. A peer reviewer can perform the same inspection from another position. The criteria do not dictate the sentence the student must write; they define what a successful explanation needs to make observable.

The research reviewed on self-assessment repeatedly points toward criteria that are concrete, task-specific, and graduated. Concrete criteria reduce the distance between an abstract quality and visible evidence. Task-specific criteria prevent a learner from substituting a global impression of ability for an assessment of the work at hand. Graduated criteria make progress detectable in steps rather than forcing a choice between “mastered” and “failed.” That matters because a learner who can see a partial improvement has more information for choosing what to practise next.

This is why a rubric, checklist, model, or script can be part of the learning mechanism. Its value is not that it gives the teacher a tidy way to assign points. Its value is that it helps a learner compare an actual performance with a standard. In examples summarized by the critical review, students examined a model, worked with criteria, marked evidence in their drafts, recorded what needed improvement, and revised. The sequence made the criteria active. A rubric left unused is only a document; a rubric used to support a judgment and a revision becomes an instrument for thinking.

There is an important design difference between a criterion and a score. A score compresses several judgments into a symbol. Compression can be useful for recording progress, but it hides the path by which the learner should improve. A criterion keeps part of that path visible. For instance, a student may award a draft “3 out of 4” for evidence, but the number alone does not show whether the problem is missing sources, weak relevance, or an unexplained connection. The next step depends on that distinction.

Criteria must also be usable by the people expected to apply them. Vague language such as “shows professionalism” or “demonstrates deep understanding” may point toward a legitimate aspiration while leaving too much interpretation unstated. More observable descriptions give the learner a better chance to detect a small change. Evidence from task-specific performance standards and from graduated standards in a motor-learning example suggests that fine-grained references can help learners notice incremental progress and adapt their strategies. The sources do not establish one ideal rubric, however. The right level of detail depends on the task, and the critical review says the field still lacks a complete account of how task type and criteria interact.

Criteria also have a motivational meaning. A standard can make a task feel controllable by showing what effort is directed toward. It can also make a judgment feel punitive if it is used only to rank or finalize students. The same checklist can support agency when it guides revision and undermine it when it functions as a hidden grading contract. Purpose therefore cannot be separated from form: a reference point becomes educationally useful when it helps the learner decide what to do, not simply when it makes a final decision look objective.

flowchart LR N01["Learning goal"] --> N02["Usable criteria or model"] N02 --> N03["Evidence from work or process"] N03 --> N04["Judgment: what meets the goal?"] N04 --> N05["Feedback: what explains the gap?"] N05 --> N06["Adjustment or revision"] N06 --> N07["Next attempt and monitoring"] N07 --> N01 N04 -. "without a next action" .-> N08["Judgment ends"]

The arrows are dependencies, not a claim that every classroom activity moves cleanly through them. If there is no learning goal, the criteria have no purpose. If there is no evidence, the judgment becomes impressionistic. If there is no interpretation, the gap is visible but unexplained. If there is no adjustment, the judgment may document learning without participating in it.

From a judgment to a changed next move

Once a learner can compare evidence with a standard, the central question changes from “What did I get?” to “What does this tell me to do?” That change is the beginning of self-regulated learning. In the research literature, self-regulated learning means setting goals and monitoring and managing thoughts, feelings, actions, resources, and context in order to reach them. Assessment contributes by generating feedback during that cycle.

The sequence has four distinguishable moments. First comes orientation: the learner understands the goal and what counts as progress. Second comes monitoring: they inspect the work or process while it is underway or after a draft is produced. Third comes interpretation: they decide what the evidence means, including why a criterion was not met or why a strategy produced a particular result. Fourth comes regulation: they revise the work, change the strategy, seek help, allocate more time, or choose a different next task. Reflection can then feed a new goal or a more informed monitoring decision.

These moments should not be confused with four separate forms to fill in. They are functions in a cycle. A self-assessment prompt can support several of them at once; a peer comment may supply external information that changes the interpretation; a teacher’s question may help the student connect a visible problem to a strategy. What matters is whether the activity helps information travel from performance to a decision.

Take a multi-step mathematics problem. A student’s initial self-assessment might be “I do not understand this.” That judgment identifies difficulty but not a useful action. With task-specific criteria, the student may notice that they selected the right equation but substituted a unit incorrectly. The next move could be to check units in each step, solve a parallel problem, and then reassess. The assessment has not merely classified the student as weak or strong. It has narrowed the problem enough for regulation to occur.

The same logic applies to a peer review. A reviewer compares a classmate’s draft with the criteria and explains one strength and one unresolved gap. The author must then interpret the comment: is the criterion genuinely unmet, is the reviewer applying it differently, or does the draft contain evidence that was not easy to find? That interpretation can lead to revision, a request for clarification, or a reasoned decision not to change the work. The peer comment is not the endpoint. It is an additional piece of evidence entering the learner’s regulatory process.

This distinction explains why “feedback” should not be treated as synonymous with information delivered by another person. Self-assessment can generate internal feedback through monitoring. Peer- and teacher-assessment can add external feedback. In either case, the learner has to connect the information to a goal and decide how to use it. The critical review describes self-generated feedback as a catalyst of self-regulation, while the broader synthesis describes seeking and providing feedback, reflecting, and calibrating judgments as learning opportunities. The source of feedback matters, but uptake matters too.

The cycle also explains why adjustment is more than a final rewrite. A learner might change the plan for studying, decide to seek help, alter the amount of effort, manage the learning environment, or choose to practise a prerequisite skill. The 2026 synthesis groups outcomes accordingly: general regulatory strategies, (meta)cognitive strategies, regulation of motivation, and regulation of resources and the environment. These categories make visible forms of learning that a single academic score can miss. They also remind us that assessment can affect what learners do around a task, not only the quality of the final product.

Affective states participate in the same loop. Confidence, motivation, satisfaction, stress, and anxiety can shape whether a learner treats a discrepancy as information or as a verdict about the self. A successful revision may strengthen a sense of competence and encourage another attempt; an opaque or humiliating judgment may reduce the willingness to inspect the work honestly. The broad review therefore analyzes affective outcomes separately from regulatory strategies while acknowledging that the two interact. Motivation and self-efficacy can be outcomes of a useful assessment process, but they can also influence the next cycle of goal-setting and effort.

There are several predictable failure paths. The first is a missing action: students are asked to rate a finished product but are not given time to revise or use the result. The second is a missing reference: criteria are so vague, global, or detached from the task that the learner cannot locate evidence. The third is a missing interpretation: a student sees that performance is below the standard but receives no help understanding why or what alternative strategy to try. The fourth is a closed social loop: peer feedback is exchanged, but the learner cannot trust the reviewer, does not understand the comment, or has no safe way to question it. In each case, the visible activity remains, while the mechanism that could make it educationally consequential is weakened.

A final failure path concerns the explanation itself. The cycle is a useful model, but the learner’s internal processing is not fully observed by the available research. The critical review calls the cognitive and affective mechanisms of self-assessment a “next black box,” and the recent reviews call for longitudinal and mediation studies. We can identify plausible links among criteria, monitoring, agency, feedback, and regulation without claiming that the links have been demonstrated in every intervention. The model should guide what to look for, not license a stronger causal claim than the evidence supports.

The practical test is therefore simple but demanding: after a self- or peer-assessment judgment, can we identify the evidence, the interpretation, and the next change? If the answer is no, the activity may still measure perception or produce a record, but it has not yet closed the learning loop. If the answer is yes, the judgment has become part of self-regulated learning—even if the first judgment needs correction and even if the eventual learning gain remains to be tested over time.

Self-assessment and peer-assessment now come into sharper focus as two routes through this same mechanism. The next question is what changes when the learner must supply and interpret the evidence alone, and what changes when another learner’s work and judgment enter the process.

References