For: Teachers looking at reflection logs, families counting study plans, learners trying to repair a setback, and leaders reading goal-setting survey results.
Short answer: Goal-setting is not one uniform educational behaviour. A pupil may set goals because they already control their learning and want to extend it. Another may write an equally detailed goal because a recent test exposed serious gaps. A third may complete an impressive planning form because the teacher required it, while continuing to study in exactly the same way. Those cases can produce the same visible activity but have different meanings. OECD’s PISA 2025 analysis explicitly discusses proactive and compensatory interpretations of goal-setting practices, including mixed links with performance. To judge whether a goal helps, trace the reason it was formed, the next decision it changes, the evidence collected and whether the learner adjusts intelligently afterward. Do not reward the quantity of plans as though every plan were proof of independent agency.
The gateway: three identical planners, three different learning problems
At a tuition table, three Secondary 2 learners have written the same target: ‘Improve my algebra accuracy to 85% within four weeks.’ They have copied a neat grid into their notebooks. Each grid lists two revision sessions a week and a practice check on Saturday. If the tutor photographs their planners, they look indistinguishable. A school that counts completed goal sheets records a 100% success rate. Yet the educational stories are radically different.
One student already solves the chapter accurately under ordinary conditions. He wants to handle unfamiliar, multi-step questions without hints. His 85% target may be too easy; he is using the form to organise extension. Another missed the prerequisite topic while ill and repeatedly confuses the order of operations. Her goal is a response to an urgent deficit. The third writes the form carefully because unfinished forms produce reminders. She has no intention of checking errors independently and may not know what an 85% score would reveal. The same observable behaviour appears in advance of a challenge, after a setback and as compliance with an administrative ritual.
The important question is not whether the students possess a planner. It is what problem their planning is trying to solve. Teachers who skip that question can unintentionally penalise the student who needs a recovery plan, praise the student whose goal is undemanding, and miss the student whose form is decorative. This is why the recent PISA finding deserves more attention than a generic instruction to ‘set SMART goals’.
The 50-second explanation
A goal is a represented future state; goal-setting is the process of choosing, checking and sometimes revising that state. Proactive goal-setting happens before a breakdown: the learner anticipates a demand and organises action. Compensatory goal-setting happens because evidence reveals a problem: the learner checks, diagnoses and adapts to regain control. Procedural goal-setting copies the appearance of planning without changing decisions. PISA 2025 reports mixed relationships between particular goal-setting behaviours and science performance; it notes that some practices may be more visible among lower-performing pupils precisely because those pupils need to check gaps or revise strategies. Therefore a high frequency of self-checking does not automatically mean strong learning, and a low frequency does not necessarily mean irresponsibility. Ask: what triggered the goal, which action changed, what did the evidence show, and what did the learner do next? Good goal-setting improves decisions rather than filling pages.
1. A target is information about desire, not proof of capacity
A student can write ‘I will explain science processes clearly’ without knowing how to construct a causal explanation. Another can write ‘I will score AL1’ without having identified the standards, current error patterns or available study time. The target is meaningful as a statement of preference or aspiration. It does not establish that the learner knows the route. Conflating goals with control is like treating a destination as evidence that a person owns a map.
A goal becomes operational when it changes selection among actions. Which question should I attempt first? Which misconception should I repair? When should I retrieve without notes? What evidence would tell me the route is too easy? What will I stop doing because it is less useful? These decisions connect the target to learning. Without them, a goal can remain emotionally impressive and practically inert.
Educators should also distinguish outcome goals from process decisions. ‘Score 80%’ describes a desired result. ‘Use three fresh questions to test whether I can solve linear equations without a worked model’ describes an action with a diagnostic purpose. Neither alone is enough for all situations. The result keeps the purpose visible; the action supplies a controllable step; the test determines whether the step worked. A well-designed goal system uses all three without pretending control over final examination marks is complete.
2. Proactive planning begins before the student has a reason to panic
Imagine a learner starting a difficult secondary mathematics unit. She checks the syllabus, notices that the next chapter depends on operations with negative numbers and tests that prerequisite on three unprompted examples. She makes one error, repairs it, and allocates a short practice session before the new chapter begins. Her goal is proactive because she used an anticipated demand to protect future learning. The trigger was not a failed examination; it was a sensible forecast of what later work would require.
Proactive planning has three characteristic advantages. It can make dependencies visible before deadlines compress choices. It can protect time for retrieval and feedback rather than relying entirely on last-minute rereading. And it can create a baseline against which the learner interprets new performance. But it is not always efficient. A student who forecasts every imaginable weakness may overplan and underlearn. The quality lies in proportionality: enough anticipation to reduce predictable failure, not an elaborate attempt to control every future uncertainty.
A teacher can help by presenting the next task’s prerequisite structure. ‘Next week we will explain rates of reaction. Here are the three ideas you need first; which can you already justify?’ This creates a real decision. It is stronger than requiring pupils to write ‘I will work harder’ because the requirement has been translated into a specific readiness condition.
3. Compensatory planning begins with a mismatch
Now consider the learner who expected to score well on a vocabulary-in-context test but misread five words because she relied on their familiar everyday meanings. She had been revising a word list accurately; the assessment required interpretation under a new context. Her plan changes after discovering the gap. She chooses contrast pairs, explains why the common meaning fails in each sentence and tests the skill on unfamiliar passages. That is compensatory planning: a response to evidence that her previous approach did not meet the actual demand.
There is nothing second-rate about this process. Recovery is part of self-regulation. In fact, refusing to change plans after failure can be a worse sign than increased goal-setting activity. Yet a school dashboard may treat more remedial goals as evidence of more agency, ignoring that the plans arose from greater difficulty. A class with many students struggling may generate more review logs than a class where the same material was already secure. Raw counts therefore cannot reveal the desired distinction.
Compensatory action becomes productive when the learner identifies a reason, adjusts a controllable practice and tests whether the specific error diminishes. It becomes ritual when the plan repeats a vague promise without isolating the failure. ‘Do more science’ is not a diagnosis. ‘In my explanation I describe what happens but omit why it happens; I will write mechanism links for three new questions and check them against marking criteria’ is closer to one.
4. Why the newest PISA evidence makes the distinction important
The OECD’s PISA 2025 Results, Volume I, published on 8 September 2026, does not treat self-regulated goal setting as a uniformly favourable personality score. Its analysis describes planning, monitoring and study adjustment as related but not interchangeable practices. Some reported goal-setting behaviours align positively with science performance on average, yet the pattern is mixed across items and systems. The chapter explicitly allows that checking understanding and revising strategies may appear more often among lower-performing learners because these actions are being used to compensate for gaps.
This is a subtle but essential reading of survey data. Suppose a question asks whether the pupil checks understanding after a lesson. A student who understands easily may check briefly and move on. A student who encounters confusing content may check repeatedly, consult examples and still perform poorly because the underlying prerequisite remains weak. The second student reports more checking. That does not mean checking causes poor results. The difficulty may be driving both the checking and the low score.
PISA associations remain observational and largely self-reported. The report does not prove that a particular recovery routine caused higher or lower performance in a specific class. We should use its distinctions to improve the questions we ask locally, not convert its statistical findings into prescriptions for individual children.
5. Stage one: identify what triggered the goal
A responsible teacher begins by asking for the first event that made the goal necessary. Was there a new syllabus demand, an approaching assessment, a missed lesson, a teacher’s feedback, an unexpected error, a change in home routine or simply a compulsory planning form? The event matters because it tells us whether the learner is preparing for an anticipated problem, repairing an existing one or responding to a rule about paperwork.
The question should not become a miniature interrogation. A student can answer in one sentence: ‘I noticed I could not explain why the algebra steps worked,’ or ‘I want to try questions without the method printed next to them.’ For young pupils, the adult may help identify the trigger from work samples. A primary learner can point to the exact sentence or calculation where the answer stopped making sense. That is already useful.
A goal with no identifiable trigger is not automatically invalid. Some projects begin from curiosity or a personally valued challenge. But if the goal is presented as an intervention for a measurable weakness, then the team needs evidence of the weakness. Otherwise the chosen practice may be unrelated to the actual barrier.
6. Stage two: name the decision the goal will change
A goal that cannot change behaviour is an aspiration, not a functioning learning plan. Ask the student to choose one decision that would have been different without the goal. For example, instead of doing twenty more familiar factorisation questions, the learner will attempt four mixed questions requiring selection between factorisation and another method. Instead of rereading ten pages, the learner will write a mechanism explanation from memory and then inspect where it fails. Instead of repeating a memorised composition phrase, the learner will compare two versions and select the one that fits the narrative situation.
The decision must be inside the learner’s reach. ‘My teacher will make the test easier’ is outside the learner’s normal control. ‘I will ask which concept my repeated error belongs to, then use a different representation for the next attempt’ is controllable. External support can be an appropriate action when it is specific and proportionate; independence never requires refusing needed help.
Keep the first change small. Multiple simultaneous changes make it difficult to tell what helped. A learner who changes tutor, textbook, revision time, practice format and sleep schedule in one week may improve, but the mechanism remains obscure. A small, credible repair can be examined and expanded later.
7. Stage three: define what would count as progress
The evidence must match the claimed learning. A planner filled with checkmarks can show that a student completed planned activities. It cannot alone show that the intended knowledge strengthened. For vocabulary, use new contexts that require meaning discrimination. For mathematics, use unfamiliar questions that require choosing a method without an obvious label. For science, ask for a causal chain rather than copied definitions. For English, require textual evidence for an inference in a fresh passage.
A useful assessment can include conditions. Was support available? Was the student timed? Did the answer rely on a nearly identical example? Was the performance repeated after a delay? These details prevent us from calling a highly scaffolded result independent mastery. They also allow partial success to be visible. A pupil who previously could not start a problem may now select the right first operation but still make an arithmetic error. That is not full mastery; it is a genuine change in the failure point.
Avoid making the outcome so narrow that it becomes self-confirming. If the goal is ‘complete four worksheets’, four completed worksheets prove compliance, not understanding. If the goal is ‘use a new reading inference reliably’, a varied independent question gives stronger evidence. The assessment should test the capability the goal was meant to build.
8. Stage four: inspect the first response to disappointing evidence
A plan is most informative when something does not work. Suppose a student completes the agreed revision sessions and the next independent check remains weak. The unhelpful response is to write the same goal in larger letters or assume insufficient effort. Instead, identify where the causal expectation failed. Was the practice too easy? Did the learner rehearse recognition rather than retrieval? Was the prerequisite still missing? Was feedback too late or too vague? Did the student understand the method but not recognise when to use it?
The goal can then be revised with a hypothesis. ‘The problem is not my ability to calculate percentages; it is identifying which base the percentage refers to.’ The new practice should discriminate this hypothesis from alternatives. A short comparison task with different denominators may be more valuable than another page of routine arithmetic. The learner learns not only content but how to react intelligently to evidence.
However, a single low result should not trigger frantic strategy switching. Performance contains noise: fatigue, unfamiliar wording, careless execution and sampling variation. Look for repeated evidence or a clearly identifiable error. The goal system should neither ignore feedback nor overreact to every mark.
9. The difference between frequency and quality
Some schools count the number of goals set, reflection entries written or strategies named. Counting is convenient. But a learner can create ten shallow goals while another makes one consequential decision and follows it through. The frequency of planning activity therefore should be a supporting indicator, not the verdict. Quantity can rise after difficulties simply because more repair is needed.
Quality is visible in the relationship between parts. Is the target based on evidence? Does the action plausibly address the gap? Is the feedback interpretable? Can the learner explain why they kept or changed the strategy? Did the action transfer to a new task? A short, coherent trace can contain more agency than a beautifully presented monthly portfolio.
A teacher may deliberately collect only one plan per pupil during a unit, then review it at three moments. This reduces administrative burden and makes the evolution of reasoning easier to see. The form should be small enough that planning does not consume the time needed for studying. A student who spends forty minutes decorating a tracker and ten minutes retrieving knowledge has probably optimised the wrong output.
10. Example: the deceptively impressive mathematics planner
Imagine Nadia, a fictional Secondary 3 learner, whose recent E-Math work shows excellent accuracy on routine simultaneous equations but repeated confusion when a word problem requires representing the unknown quantities. Her goal sheet initially says ‘Do more simultaneous equations, 30 minutes every day.’ This is a common mismatch: the plan targets a familiar operation instead of the difficult decision that the assessment exposed.
A tutor asks her to show the exact point where each failed question broke. Nadia notices that she often writes correct algebra after choosing the wrong variables. She changes the goal to ‘Before solving, write a one-sentence definition of each variable and test it against every relation in the question.’ The daily practice becomes two contrasting problems rather than thirty minutes of undifferentiated calculation. On a later unseen problem, she defines variables correctly without a prompt but still miscopies one number. The result is partial: representation has improved; execution needs attention.
Her planner now contains less activity but more learning control. The improvement is not the neatness of the goal or the number of checks. It is the tighter fit between evidence, chosen action and the next independent performance.
11. Example: the English essay target that rewards the wrong action
A Secondary 4 student receives feedback that his essays are ‘too descriptive’. He sets a target to write one thousand words every evening. Word count is observable and easy to verify. It is also poorly aligned with the concern. His essays could grow longer while remaining descriptive. The missing capability is selecting and explaining evidence that supports an argument.
A teacher asks him to take one existing paragraph, highlight each claim, underline supporting evidence and identify the reasoning that connects them. He discovers that some paragraphs contain elegant detail without an answer to the question. His revised goal is to produce three analytical paragraphs a week, each with a warranted inference and one consideration of an alternative reading. He checks the skill on fresh prompts and later under examination timing.
The distinction between proactive and compensatory planning appears here too. The student is repairing a known weakness. A lower mark is not evidence that planning was harmful; it helps explain why planning became necessary. The relevant outcome is whether his argument improves, not whether he can report more study sessions than his classmates.
12. Example: the primary learner who mistakes correction for mastery
A Primary 5 pupil has a page of corrected science answers. Every answer is now right because the teacher has written explanations beside the mistakes. The pupil promises, ‘Next time I will check my work more carefully.’ A parent sees a responsible intention. The missing question is what checking could have detected. The pupil often omits causal links because she recognises key terms but has not learned how to explain the mechanism.
Instead of simply checking for spelling or missing words, the pupil is taught to mark each cause, intermediate process and outcome in a model explanation. She writes a fresh explanation without notes, then checks whether the causal chain is complete. A week later, an unfamiliar question tests the same reasoning demand. If she still cannot start, the goal returns to direct instruction and model comparison rather than more self-checking.
The example illustrates a boundary: goal-setting is not a substitute for teaching a skill the student has not acquired. Good self-regulation sometimes means recognising that a prerequisite or model is missing and requesting it.
13. Example: when high-achieving students also misuse goals
A strong student can set goals badly. Imagine Jun, already accurate on most routine questions, selecting ‘achieve at least 75%’ as a target because it is safely below his current level. The goal looks realistic and measurable, but it will not distinguish further learning from maintenance. He may avoid unfamiliar problems and still achieve the target comfortably. The planner records success while concealing the opportunity cost of underchallenge.
His tutor changes the target from a safe percentage to a capability: solve a new problem without the method signposted, explain why the chosen approach fits and evaluate an alternative solution. Timed practice can be added later, but the immediate question is independent strategy selection. Jun now encounters productive uncertainty. His first attempt may be slower and less polished than his rehearsed work. That temporary drop is not necessarily failure if it comes with genuine extension and later independent recovery.
A uniform classroom goal sheet often misses this problem because it assumes goals are needed only to repair weakness. Proactive goals can protect meaningful challenge for already secure learners, but only if the target is demanding enough to change their decisions.
14. What ‘agency’ means at the classroom floor
Agency is sometimes presented as the freedom to choose whatever one wants. For learning, that is too weak. A pupil may choose an attractive worksheet because it feels easy. Another may choose a hard task that is beyond current prerequisites. Neither choice automatically improves capability. Educational agency requires enough information and skill to choose responsibly among actions, monitor the result and revise when necessary.
A teacher can grant bounded choice: two problem types that target the same concept, alternative evidence sources with comparable quality or a choice between explaining a diagram and writing a causal paragraph. Explain the trade-offs. Then ask the student to justify the choice and assess the resulting work. This is not control surrendered by the teacher; it is control deliberately transferred in a way the learner can use.
The transfer should be gradual. An early learner may need the teacher to identify the target and select the first practice. Later, the learner can choose from two strategies, then propose a justified strategy, then identify and plan an unfamiliar learning problem independently. Agency is developed through these decisions, not declared by a poster saying ‘Take ownership’.
15. What parents should do with weekly planners
When a child brings home a goal-setting sheet, parents can ask three conversational questions rather than scoring its appearance. ‘What made you choose that goal?’ reveals the trigger. ‘What will you do differently tomorrow?’ reveals the action. ‘How will you know if that helped?’ reveals the evidence. If the child cannot answer, the parent can help narrow the task without rewriting the entire planner on the child’s behalf.
Families should be careful not to convert a repair goal into moral evidence. A pupil who creates several catch-up plans may have missed lessons through illness, experienced an unusually difficult unit or recognised a weakness promptly. The existence of the plan is not evidence of poor character. The relevant concern is whether barriers are being addressed and whether the learner has access to appropriate support.
Similarly, parents should not reward ambitious targets without checking plausibility. ‘Get full marks in every subject’ may create anxiety without indicating a pathway. A target should respect time, sleep, available knowledge, school obligations and the child’s real circumstances. If a serious wellbeing or access issue is limiting study, that issue needs appropriate adults and support; another goal sheet cannot replace them.
16. How teachers can reduce compliance theatre
A planning routine becomes theatre when adults inspect its formatting more carefully than the decision it produces. Notice the warning signs: identical goals across students with different errors, reflection sentences repeated every week, targets chosen only because the template offers them, and pupils who can recite their goal but cannot show one changed attempt. This is not evidence that goal-setting is useless. It shows that the administrative representation has drifted away from the mechanism.
Reduce the form to four fields: trigger, action, evidence, next decision. Allow a learner to say ‘I will keep the strategy because the last three fresh checks improved’ as well as ‘I will switch because the same error remains.’ Both are forms of regulation. Do not force every student to announce a new strategy each week just to show activity. Sometimes the intelligent choice is to continue a working routine.
Use student work to audit claims. A teacher who sees ‘I understand photosynthesis now’ can ask for an explanation of energy transfer and matter in a fresh setting. A parent who sees ‘I have improved vocabulary’ can ask for accurate use of a word in context. The goal becomes a hypothesis to test rather than a sentence to praise.
17. An evidence-aware four-week implementation route
In week one, collect one small baseline task and ask learners to identify a specific opportunity or difficulty. Record who chose the goal and whether the trigger was a future challenge, current gap or compulsory exercise. Do not compare students’ counts as a proxy for motivation. Protect privacy and avoid attaching deficiency labels to the child.
In week two, teach one example of linking evidence to action. Demonstrate how a vague target becomes a discriminating practice: ‘Improve English’ becomes ‘Use context to distinguish a word’s literal and implied meanings in three fresh passages.’ Give feedback on the causal fit, not only the wording of the goal. In week three, perform a check on new material and let students choose to continue, modify or seek targeted instruction. Require a reason, not a change for its own sake.
In week four, inspect whether learners’ decisions became more accurate. Did students identify the correct failure point? Did their practice align with that point? Did they recognise when they needed teaching rather than more self-monitoring? Did successful pupils choose real extension instead of safe targets? Count these improvements before celebrating a rise in completed forms. A four-week exercise is a local teaching test, not proof of a general causal effect.
18. Evidence, uncertainty and the limits of a neat story
The OECD’s findings are a useful challenge to simplistic goal-setting advocacy. But it would be equally simplistic to claim that compensatory planning is the explanation for every negative association between an item and performance. Different tasks, cultures, demands, response interpretations and learner histories can produce different patterns. Performance may also influence self-report: after a poor result, a student may become more aware of needing to check understanding. A high performer may underestimate the planning they already do automatically.
The EEF’s metacognition and self-regulated learning guidance supports explicit teaching of planning, monitoring and evaluation as part of subject learning, rather than assuming learners can self-regulate because a worksheet asks them to reflect. The What Works Clearinghouse guidance on studying offers complementary instruction principles for practice and retrieval. Neither provides a universal ratio of ‘proactive’ to ‘compensatory’ goals for every Singapore classroom. That ratio would require local definition, measurement and evidence.
A goal is useful when it leads to better diagnosis and decisions. Sometimes that means more planning; sometimes it means shorter planning and deeper practice. Do not confuse visible paperwork with a learner’s capacity to act.
19. Common failure modes and repairs
Rewarding more plans. The pupil who struggles most may create the most recovery goals. Reward decision quality, not the absolute count.
Assuming a negative association implies harm. Poorer performers may use self-checks because they are having difficulty. Preserve the order of events and test plausible confounders.
Overreacting after one wrong answer. A single mistake can be noise. Use repeated, interpretable evidence or a clearly diagnosed conceptual failure before changing the method.
Making goals too easy for successful pupils. A comfortable target can conceal underchallenge. Ask for transfer and independent selection where core fluency is secure.
Writing goals instead of teaching prerequisites. A pupil cannot self-regulate a concept they have never understood into existence. Provide models, explanation and adequate instruction.
Treating a plan as a promise of a final mark. The student controls practice decisions only partly controls results. Give realistic targets and review conditions.
Turning reflection into public performance. Learners may write what sounds acceptable. Use private, low-stakes evidence and a conversation about actual work.
Frequently asked questions
Is goal-setting good or bad for students?
It can help when it converts a meaningful target into action, monitoring and adjustment. It can become a distraction when it replaces learning or rewards paperwork. The mechanism and context matter more than the label.
Why can struggling pupils report more self-checking than successful pupils?
They may need to detect errors, check understanding and revise strategies more often. This does not mean checking causes failure. Difficulty can be a common driver of both greater checking and lower attainment.
What is compensatory goal-setting?
It is planning or monitoring triggered by a recognised gap or setback. A learner uses evidence to regain control: find the mistake, choose a repair, check the next attempt and adapt again if needed.
What is proactive goal-setting?
It is planning ahead of an anticipated learning demand. The student checks prerequisites, makes a reasonable forecast, prepares a strategy and leaves room to update it as new information arrives.
Does every good goal have to be numerical?
No. A specific capability can be more meaningful than a percentage. ‘Explain the first cause-and-effect link without notes on a fresh task’ is assessable even if the measure is not a single score.
Is an AL1 goal enough for Primary 6?
It is an aspiration, not a practice programme. Pair it with current evidence about particular errors, a suitable learning route, independent checks and appropriate preparation for the assessment format.
Should a student change methods every time a goal is missed?
No. Review the size and pattern of the miss, the quality of the evidence and the likely cause. Some useful methods require sustained practice. Change the method when the evidence supports a reason, not when discomfort appears.
Can teachers set goals for children?
Yes, especially when pupils need an initial model of what a feasible goal means. Over time, explain the reasoning and transfer decisions to students according to readiness. Forced instant independence may be counterproductive.
How many goals should a pupil set at once?
There is no universal number. Start with few enough that the learner can remember, practise and evaluate them without displacing subject learning, rest and other obligations.
Are reflection forms a good measurement of agency?
They may show whether the pupil can articulate a plan under those conditions. To test agency, also observe a new decision: does the student choose, monitor and adapt a method when the template is absent?
What can a parent do after an unsuccessful goal?
Ask what the learner tried, what feedback was available, and which part of the plan failed. Support a specific next action or appropriate school help rather than interpreting the missed target as a fixed personal trait.
What if the student has no spare study time?
A plan must acknowledge real constraints. Improving a learning route may mean replacing low-value practice, adjusting load or coordinating school support, not adding hours the family does not have.
Surgical internal links
See How Education Works and the How X Works hub for the broader learning system. The general mechanism of setting a worthwhile target belongs to How Learning Goals Work. For the separate problem of converting plans into actions, see How Studying Works: The Intention–Behavior Gap. For practical lesson assessment, How Progress Tracking Works covers measuring change without mistaking forms for progress. The Clementi Secondary 1 Mathematics reference is the local tutorial standard behind the mathematics examples.
Sources and editorial caveats
- OECD (8 September 2026), PISA 2025 Results (Volume I), Student attitudes towards learning. Especially its caution about item-specific goal-setting behaviours and potentially compensatory uses.
- Education Endowment Foundation, Metacognition and Self-regulated Learning. Guidance for embedding planning, monitoring and evaluation within subject teaching.
- What Works Clearinghouse, Organizing Instruction and Study to Improve Student Learning. Practical evidence-informed design of practice and learning checks.
The students and classrooms in this article are fictional teaching cases. Percentages are illustrative targets, not asserted outcomes. The article interprets the OECD reporting cautiously; it does not claim that any individual’s motivation or learning history is recoverable from one questionnaire response.