HSW-0236
The student with the neatest file may not know how to study.
The student with the highest marks may not know how to recover when a familiar method stops working. The quiet student who submits everything on time may be using strong self-regulation—or may simply be succeeding inside a tightly structured environment that does most of the regulating. The untidy student who occasionally forgets a worksheet may nevertheless understand exactly when to retrieve, monitor comprehension, change strategy and protect attention.
Adults must make judgments anyway. Teachers decide who needs support. Tutors decide whether to reteach a strategy. Parents decide whether a child can manage revision independently. But self-regulated learning is partly hidden. What is visible—grades, handwriting, compliance, speed, participation, confidence—can become a substitute for what we actually want to know.
Two 2026 studies in Metacognition and Learning sharpen this problem. One examined how accurately teachers judged students’ knowledge of self-regulated learning strategies and found low accuracy alongside effects associated with student achievement and gender. A second experimentally manipulated the information available to teachers and found that adding non-diagnostic cues reduced judgment accuracy, even when teachers were warned that those cues were not useful.
This article calls the educational problem the self-regulation judgment blind spot: the gap between the learner’s actual strategy knowledge and use, and the impression an observer forms from easier-to-see cues.
The direct answer
Grades and good behaviour can provide useful information about a student, but they do not directly reveal how that student plans, monitors, selects strategies, regulates motivation, manages time or adapts after failure. Judgments become more accurate when adults use evidence that is closer to the self-regulatory process itself: what strategy the learner would choose, why, under which conditions, how they monitor it, and whether they can adapt when the first plan fails.
The important move is not to distrust teachers, tutors or parents. It is to improve the evidence interface. Self-regulation is difficult to infer because much of it happens before, between and after the visible actions that classrooms naturally record.
This gives HSW-0236 a specific job. It is not a replacement for the broader eduKateSengkang guide How Diagnostic Teaching Works, which explains the general logic of inference from learning evidence. HSW-0236 focuses on a narrower question: what goes wrong when observers infer a student’s self-regulated learning from cues that are easier to see than the strategies themselves?
Self-regulation is not one visible behaviour
Self-regulated learning is often compressed into a personality description: organised, independent, motivated, disciplined. That compression is convenient and frequently misleading.
A learner can be strong in one regulatory component and weak in another. They may schedule time well but use low-yield study methods. They may know effective strategies but fail to deploy them under pressure. They may monitor comprehension accurately yet struggle to initiate work. They may regulate motivation well in Mathematics and poorly in writing. They may know what a good revision method is without using it consistently.
That means at least five things should be kept separate:
- Achievement: what the learner currently scores or can perform.
- Visible behaviour: punctuality, neatness, participation, persistence and compliance.
- Strategy knowledge: whether the learner knows which strategies are useful, when and why.
- Strategy use: whether the learner actually deploys those strategies in real study.
- Strategy quality: whether the strategy is enacted well enough to change learning.
These layers interact, but none is a perfect proxy for another. A high score may reflect strong prior knowledge, intensive external support, a familiar task or good self-regulation. A low score may reflect weak subject knowledge even when regulation is thoughtful. A neat book may reflect care—or a preference for presentation that consumes time without improving memory.
What the first 2026 study found
Sabrina Nadja Brunner and Yves Karlen studied teacher judgments of students’ self-regulated learning strategy knowledge across 200 students, 69 teachers and 16 schools in Switzerland and Germany, with most students in Grades 7 and 8. The student measures examined knowledge across areas including cognitive and metacognitive strategies, time management, learning-environment management and motivation regulation.
The central finding was uncomfortable but useful: teachers’ judgments showed low accuracy and weak differentiation across components of self-regulated learning. Judgments were also associated with student characteristics that were not equivalent to the specific strategy knowledge being estimated, including achievement and gender. The authors discuss these patterns in terms of global impressions and halo-like effects.
That does not mean teachers are careless. It means the target is hard to observe. Students do much of their planning, monitoring and strategy selection outside the moments a teacher can see. Teachers therefore face the ordinary human problem of filling evidence gaps with salient cues.
The study also found that teacher characteristics such as self-reported knowledge, self-concept and experience did not straightforwardly eliminate the judgment biases. That result should not be overgeneralised into “training cannot help.” It shows that possessing general knowledge or experience is not by itself enough to guarantee accurate inference about a particular learner’s self-regulatory strategy knowledge.
What the second 2026 study added
A separate 2026 study on teachers’ assessment competence in self-regulated learning moved from naturally occurring judgments to an experimental cue-manipulation design. The researchers worked with more than 260 pre-service and in-service secondary teachers across Switzerland, Turkey and Germany.
Participants received information designed to support judgments about learners’ self-regulated learning. In one condition, they saw diagnostic cues. In another, diagnostic information was accompanied by non-diagnostic cues, including information such as handwriting and social school-report features. A third condition added an explicit warning that these additional cues had low diagnostic value.
Judgment accuracy was highest when diagnostic cues were presented without the distracting non-diagnostic information. Adding non-diagnostic cues reduced accuracy. Warning participants about the low diagnostic value of those cues did not fully protect judgment.
This is a powerful result because it separates two explanations. Poor judgment is not only a problem of missing information. It can also arise when useful information is present but is accompanied by vivid, easy-to-process cues that should receive little weight.
Again, the boundary matters. This was a designed judgment task, not proof that every teacher in a live classroom will be misled by handwriting or school-report information in the same way. It provides experimental evidence about cue use under the study conditions.
The proxy problem: what is easy to see starts standing in for what matters
Classrooms produce many visible signals.
- Who submits on time.
- Who looks attentive.
- Whose notes are neat.
- Who asks for help.
- Who gets high marks.
- Who works quickly.
- Who remembers materials.
- Who appears confident.
Each cue can contain real information. The problem begins when an observer silently upgrades it into evidence for a different construct.
“She gets A grades” becomes “she knows how to regulate learning.”
“He needs reminders” becomes “he does not understand study strategies.”
“Her notes are beautiful” becomes “her learning system is organised.”
“He studies for a long time” becomes “he knows how to allocate effort.”
These inferences may sometimes be right. What makes them dangerous is not that they are always false, but that the evidence is indirect and can support several competing explanations.
Illustrative case: two students who look easy to classify
The following case is illustrative, not research data.
Clara has immaculate notes, completes every worksheet and scores 82% on familiar tests. Her teacher describes her as highly self-regulated. During a diagnostic conversation, however, Clara explains that when she revises she rereads until the page feels familiar. She cannot say how she decides what is secure, rarely tests herself without notes and increases study time rather than changing method when performance stalls.
Rafi’s folder is messy. He occasionally forgets to file a worksheet and scores 72%. His visible organisation looks weaker. Yet when given a new examination topic, he can identify what he knows, choose retrieval for factual content, switch to worked examples when a procedure is unstable, compare his answer against a criterion and schedule a delayed check on the weak area.
It would be wrong to conclude that Rafi is globally better at self-regulation or that Clara is globally poor. Clara’s strong habits may support real learning; Rafi’s material organisation may still create avoidable costs. The point is narrower: surface impressions are insufficient to infer the specific regulatory capability being diagnosed.
Diagnostic cues versus non-diagnostic cues
A cue is diagnostic when it has a defensible relationship to the question we are trying to answer.
If the question is, “Does this learner know when retrieval practice is useful?”, asking the learner to compare retrieval with rereading in a realistic study scenario is more diagnostic than inspecting handwriting. If the question is, “Can this learner monitor comprehension?”, a delayed prediction followed by performance evidence is more diagnostic than asking whether the learner feels generally confident. If the question is, “Can this learner regulate time?”, examining how they plan and adapt a week with competing demands is more diagnostic than observing whether one homework task arrived on time.
Non-diagnostic does not mean worthless. Handwriting may matter for legibility. Punctuality may matter for responsibility. Participation may matter for classroom interaction. The mistake is using a cue to answer a question it was not built to answer.
Strategy knowledge is not strategy use
The 2026 teacher-judgment study focused substantially on students’ strategy knowledge. That distinction deserves protection.
A learner may know that self-testing is more diagnostic than rereading and still reread because testing feels difficult. A learner may know that distraction should be controlled and still study beside a phone because social rewards are immediate. A learner may understand planning principles and fail to start because the first task is emotionally aversive or poorly specified.
Conversely, a learner can sometimes use a productive routine without being able to explain its theoretical rationale. Habit, teacher structure or imitation may produce competent behaviour before explicit strategy knowledge is fully articulated.
Therefore a complete diagnosis asks at least three separate questions:
- Does the learner know an appropriate strategy?
- Does the learner choose and use it under realistic conditions?
- Does that use improve delayed or independent performance?
This connects to HSW-0218 on strategy enactment quality, which owns the distinction between choosing a useful method and carrying it out well.
How to make self-regulation more observable without pretending to read the mind
The goal is not surveillance. It is to create small, relevant windows into the learner’s decision process.
Use strategy-choice scenarios
Present a concrete study problem: “You can explain the chapter while looking at the notes, but tomorrow you must answer without them. You have twenty minutes left. What do you do next, and why?” The learner’s answer reveals conditional strategy knowledge more directly than a general question such as “Are you good at studying?”
Ask for the trigger
“How do you know when to stop rereading?” “What tells you that a topic deserves another session?” “When do you switch from a worked example to independent practice?” Triggers reveal how the learner connects monitoring to control.
Observe a real transition
Watch what happens when a familiar method fails. Does the learner repeat harder, seek an answer immediately, change representation, retrieve a prerequisite, ask a discriminating question or abandon the task? Strategy flexibility becomes visible at transition points.
Compare prediction with later evidence
Ask the learner to predict which items will be remembered, then test later. This does not measure all of self-regulation, but it provides evidence about monitoring accuracy and whether the learner updates study choices when predictions fail.
Use learning traces cautiously
A timetable, error log or revision history can reveal behaviour, but the trace does not explain itself. A crossed-off task could mean mastery, abandonment or superficial completion. Ask the learner to connect the trace to the decision.
The danger of the global learner label
Once adults form a global judgment—“independent”, “disorganised”, “mature”, “lazy”, “strategic”—new evidence is easily interpreted through the label.
A missed deadline by an “independent” student becomes an exception. The same missed deadline by a “disorganised” student becomes confirmation. A high mark can cause adults to overlook inefficient learning methods because the outcome appears to validate the process. A lower mark can cause strong regulatory decisions to disappear from view because achievement dominates the impression.
The 2026 findings on achievement-related judgment bias make this especially important. Achievement is educationally meaningful, but it should not be allowed to answer every question about how the student learns.
The repair is to make judgments component-specific and revisable. Instead of “Aisha is highly self-regulated,” write a claim that can be tested: “Aisha can identify when retrieval is needed and can plan her revision blocks, but currently needs prompts to switch strategy when practice reveals a prerequisite gap.”
The second statement is less elegant. It is also much more useful.
For parents: do not confuse being easy to manage with learning to manage yourself
A child can be easy to manage because the household provides excellent regulation: fixed homework time, reminders, prepared materials, quiet space, device limits and adults who notice deadlines. These structures can be beneficial. But they may hide how much control belongs to the environment.
Independence becomes visible when one support is removed and the learner compensates. If the reminder disappears, does the child notice the deadline? If the worksheet does not specify the method, can the child choose one? If the first plan fails, can the child replan rather than waiting for rescue?
Parents should therefore avoid two opposite errors. Do not remove all support to “test independence” before the learner is ready. But do not mistake permanent support for acquired self-regulation either.
A useful question is: which part of this control can the learner carry next?
For tutors and teachers: diagnose the strategy, not the stereotype
When a learner struggles, resist the urge to begin with a trait label. Begin with a task-linked question.
- What did the learner plan to do?
- What strategy did they choose?
- What cue drove that choice?
- What evidence did they monitor?
- What happened when the plan stopped working?
- Which part could the learner explain independently?
- Did the strategy survive a delay or changed task?
This approach does not eliminate judgment. Teaching requires judgment. It simply moves the judgment closer to the mechanism.
The same evidence discipline appears in the sibling-site guide How Metacognition Works in Learning: monitoring and control need observable evidence and opportunities for correction rather than assumptions about what the learner must be thinking.
A compact diagnostic interview
For one weak topic, ask the learner:
- What do you think the problem is?
- What would you do first if nobody gave you a study method?
- Why that method?
- What would tell you it is not working?
- What would you switch to?
- How will you know tomorrow whether the repair held?
This six-question conversation cannot diagnose all self-regulation. It can, however, reveal whether the learner has conditional strategy knowledge, monitoring criteria and an adaptation plan. It also reduces the need to infer those processes from neatness or grades.
Delayed and independent performance checks
Because self-regulation is partly about managing future learning, the strongest evidence often appears after support is withdrawn.
Teach or discuss a strategy. Then wait. Present a fresh problem without naming the strategy. Change one surface feature. Ask the learner to plan, select a method, monitor progress and explain any switch. Finally, inspect the quality of the resulting performance.
If the learner can describe the strategy but never selects it, strategy knowledge is not the first problem. If the learner selects it but executes it poorly, the problem is enactment. If the learner performs well only while an adult structures every transition, independent regulation has not yet been demonstrated. If the learner adapts appropriately after a mismatch, that is stronger evidence than simply following a rehearsed plan.
Delayed performance is especially valuable because visible classroom compliance can temporarily support behaviour. A later independent task asks whether the control has become portable.
Misconceptions
“Grades tell us nothing about self-regulation.” Too strong. Achievement can be informative and was associated with teachers’ judgments in the 2026 work. The problem is treating achievement as direct measurement of specific regulatory components.
“Neatness and punctuality are irrelevant.” No. They matter for some educational outcomes. They become non-diagnostic when used to answer a different question, such as whether the learner knows how to monitor comprehension.
“Teachers are biased, so student self-report is better.” Student self-report has its own limitations. Learners can misremember, overgeneralise or describe what they believe they should do rather than what they actually do. Strong diagnosis triangulates sources.
“A warning about irrelevant cues fixes the problem.” The 2026 experimental study suggests that merely warning assessors may not fully protect accuracy when non-diagnostic cues remain present.
“Strategy knowledge is the whole of self-regulation.” It is not. Knowledge, use, quality, motivation, monitoring and adaptation must be distinguished.
Evidence and limits
The first primary source is Brunner and Karlen, “Teacher judgments of student self-regulated learning – Examining the effect of student and teacher characteristics”, published in Metacognition and Learning on 26 May 2026 (DOI: 10.1007/s11409-026-09469-x). The study compared teachers’ judgments with students’ tested strategy knowledge across 200 students, 69 teachers and 16 schools in Switzerland and Germany. It found low judgment accuracy and component specificity, with achievement- and gender-related effects in judgments.
The second primary source is “Teachers’ assessment competence in self-regulated learning – presence of non-diagnostic cues diminishes judgment accuracy”, published in Metacognition and Learning on 21 April 2026 (DOI: 10.1007/s11409-026-09455-3). In an experimental judgment task with more than 260 pre-service and in-service secondary teachers from Switzerland, Turkey and Germany, accuracy was higher when diagnostic information was presented without additional non-diagnostic cues; warnings about cue diagnosticity did not fully remove the effect.
Important limits remain. Strategy-knowledge tests are not a complete measure of real-world self-regulated learning. The school samples and national contexts limit generalisation. The experimental cue study simplifies the information environment compared with live classrooms. Neither paper supports diagnosing an individual learner from group averages, and neither justifies treating observed bias patterns as fixed traits in teachers or students.
For broader practical guidance on teaching metacognitive and self-regulatory strategies, the Education Endowment Foundation’s second-edition Metacognition and Self-Regulated Learning guidance emphasises explicit strategy teaching, modelling, guided practice and movement toward independence. That evidence supports making regulation teachable and inspectable; it does not remove the need for accurate learner-specific diagnosis.
The return
The difficulty is not that self-regulation is invisible. It is that it is only partly visible, and the classroom offers many easier signals that compete for attention.
Grades matter. Behaviour matters. Organisation matters. But none should be allowed to impersonate the whole learning control system.
A more humane and rigorous judgment stays close to the evidence: what the learner knows about strategies, what they choose, what they actually do, what they notice when the method fails, and what survives when the support is gone.
That turns “She is a good independent learner” from a flattering label into a question we can investigate. And it turns “He is disorganised” from a verdict into a set of components we can teach.
The point of diagnosis is not to classify the student more confidently. It is to find the next control the learner can acquire.
Continue through the How Studying Works Numbered Series Reading Index, or return to the How X Works Hub.