The recipe is on the table. A Primary 5 student in Clementi is conducting a simple home science activity with permission and adult supervision: comparing how quickly sugar dissolves in water under different stirring conditions. She has read the instructions. She knows the words. She measures the water, adds sugar, reaches for a stopwatch, and hears somebody calling from the next room. Thirty seconds later she returns. Has she stirred the first cup already? Did she start the timer before or after stirring? Is the second cup waiting for sugar, or did she add it while distracted?
The science has not suddenly become more advanced. What has disappeared is the student’s reliable knowledge of the current state of the procedure. A list of steps tells her what to do in general. It does not necessarily tell her which steps have been completed in this particular attempt, what conditions now hold, and where it is safe to continue. A child may follow each isolated instruction correctly yet lose the experiment when the sequence is interrupted.
The same thing happens in Mathematics when a pupil returns to a partly solved equation and repeats a subtraction, in English when a learner edits the same paragraph twice while leaving another untouched, and in project work when a student cannot remember which source has already been checked. Parents searching for why my child forgets instructions halfway, how to follow multi-step procedures, students losing place in science experiments, task interruptions in homework or how to restart after distraction may be seeing one narrow, repairable mechanism.
The 50-second answer
A multi-step task requires more than remembering the instructions. The learner also needs to know the state of the work: what was done, what changed, what remains and what must be checked before continuing. An interruption can break this connection even when the student understands every individual step. A practical re-entry checkpoint records three things: last verified completed step, current observable state, and next safe action. When returning, the learner checks the record against the actual work before resuming. This should be taught within real subject tasks and gradually simplified as competence develops; it is not a reason to give every child an enormous checklist. Research on interruptions and metacognition supports a plausible rationale for external cues, but the precise checkpoint format here is a practical teaching design, not a universally validated clinical method.
Knowing the list is not the same as knowing where you are
Imagine giving a child a perfect list of ten instructions. At the start, the list is helpful. At step six, the child may still know all ten instructions yet be uncertain whether step five happened. The problem is no longer understanding the list; it is matching the plan to the current material state. In a cooking activity, ingredients that have been added cannot simply be “added again” because the learner forgot doing so. In Mathematics, an operation applied twice changes the answer. In a Science investigation, a measurement taken before the timer starts is not equivalent to one taken afterward.
This makes procedures different from ordinary reading. A story remains on the page when the reader is interrupted. A process changes the world as it is performed. The beaker contains a new mixture, the equation has been transformed, the draft contains revised sentences, the file has been saved or not saved. The learner must reconstruct both the abstract instruction and the changed state of the task. The most reliable question on return is not simply “What comes next?” It is “What do I know has already happened?”
Many school tasks conceal this distinction because adults can reconstruct state easily. A teacher glances at a half-solved equation and recognises the last valid transformation. A child with weaker familiarity sees a page of symbols and remembers only that they were “doing algebra.” The difference may be domain expertise, not general obedience. Teaching re-entry as a subject-specific skill makes the invisible part of the task visible.
Stage one: define the unit of work that can actually finish
A step should end with something observable. “Understand the experiment” is not a completed procedural step. “Label the cups A and B” is. “Do the Mathematics” is too broad. “Substitute x = 4 into the formula and simplify the numerator” is a step with a checkable result. Good procedures connect actions to states: after you have done X, you should be able to see or verify Y.
This matters because students often use vague ticks. A child may check off “write introduction” after typing two unfinished sentences. A Science student may tick “measure water” despite not recording the volume. The checklist gives an appearance of order while the underlying work remains incomplete. The repair is to identify a small evidence marker for each meaningful transition. “100 mL measured in Cup A and recorded” is more useful than “water done.”
Not every task requires such precision. A familiar routine like sharpening a pencil does not need a formal record. The checkpoint should be reserved for tasks in which skipping, duplicating or misordering a step would matter. This is especially valuable when experiments have several conditions, calculations have dependent transformations, or a written response must preserve evidence through multiple revisions.
Stage two: mark the difference between an action and a verified result
A student can perform an action without achieving the intended state. They may click Save but the file may not finish uploading. They may copy a number but miss the minus sign. They may pour a measured quantity but spill a little. The right checkpoint confirms the result rather than simply remembering the action. This is the difference between “I pressed the button” and “The screen showed Saved at 4:10”; between “I subtracted” and “Both sides of the equation are still balanced.”
For a Maths problem, the checkpoint might be 2x+6=18 → subtract 6 from both sides → 2x=12. The learner can verify this transformation before continuing to divide by two. If distracted, she does not need to trust an uncertain memory about whether six was subtracted. She can read the equation in its current state and compare it with the recorded step. The page becomes an external witness to what happened.
In a Science practical, a checkpoint might be Cup A: 100 mL water, 5 g sugar, timer not started. These values must correspond to what can still be verified. If an irreversible action has occurred and its timing is uncertain, the correct re-entry may be to restart that trial rather than invent a measurement. Scientific integrity sometimes requires discarding an ambiguous run. The procedure should explicitly allow that option.
Stage three: make the next action conditional on current state
Ordinary instructions often read as a chain: do A, then B, then C. Re-entry requires a branch: if A is confirmed complete, proceed to B; if A is uncertain and reversible, verify or safely repeat it; if A is irreversible and uncertain, stop and ask whether the trial must restart. This conditional structure is the heart of recovery. It prevents a student treating a sequence as a spell to be recited from wherever memory happens to resume.
Consider an English editing task with three passes: argument, paragraph structure, sentence accuracy. The learner has checked the first two sections for argument and is interrupted. On return, “continue editing” is ambiguous. Should she resume with paragraph three’s argument, start the structure pass, or reread everything? A checkpoint stating argument pass done: paragraphs 1–2; next: paragraph 3 supplies a testable route. It protects time without telling the learner which argument is good.
In Mathematics, a child can add a small mark beside the last equation known to be valid. If the next line is doubtful, return to the last verified line rather than repeating the entire solution automatically. When teaching, distinguish “last line written” from “last line checked.” They are not necessarily the same. This difference prevents students building a long solution on top of an unnoticed error.
Stage four: teach a pause that costs seconds rather than minutes
Some children hear “checkpoints” and imagine another page of administrative writing. That would defeat the purpose. For many tasks, the checkpoint can be three short fragments: DONE — STATE — NEXT. In a science table: A measured — 100 mL / 5 g — start timing once stirring begins. In a composition: claims checked to para 2 — para 3 unreviewed — resume at topic sentence. In algebra: collected terms — 3x=15 verified — divide by 3.
The marker should be placed where the work lives. A note on the other side of the room will be ignored. A separate digital document may require so much switching that it introduces a new interruption. A sticky note, a margin tick, a line in a notebook or a comment inside the draft can be enough. If the task is electronic, use the platform’s available save and version features rather than assuming every app autosaves reliably.
A few seconds spent making a precise checkpoint can save repeated rereading, but not every pause deserves one. The sensible trigger is an impending interruption, the end of a long segment, or a transition where the next action would be costly if performed twice. Teach students to recognise those moments. Over time, a competent learner can make more of these decisions mentally without writing every micro-step.
Stage five: rehearse safe re-entry, not just the beginning
Teachers commonly model a procedure from step one to completion. Students watch the smooth demonstration, then try to reproduce it. Real work is less smooth. Devices interrupt, peers ask questions, a measurement has to be repeated, or a child loses concentration. A robust lesson includes one deliberately staged pause: stop after a middle step, move briefly to another harmless task, then return and narrate the recovery.
The teacher might say, “I am back at the equation. I cannot assume the line I last wrote is correct. I can verify that subtracting six from both sides produced 2x = 12. My next action is dividing by two.” This models a decision, not just a sequence. Students see that expert practice contains checks, not merely speed. In Science, a demonstration can show how to inspect a labelled cup and log before resuming, or when uncertainty means the run should be discarded.
A rehearsal should also include an honest failure. The teacher might discover that a measurement was never recorded and say, “I cannot recover that value from memory accurately. I need to redo this trial.” This is a stronger scientific lesson than improvising a plausible number. The goal is an operational habit: re-enter from verified state, not from wishful memory.
The interruption has a cognitive cost, but not every cost is identical
Research on interruptions provides a useful foundation for the problem. A 2024 study reported in Scientific Reports investigated how interruptions affect working-memory tasks across age groups. Its details concern laboratory tasks and adults, not Primary 5 sugar experiments, so it should not be treated as direct evidence that a specific school checklist raises marks. It does show that interrupted tasks can impose demands on refocusing and recovery. Other work on resumption costs in Memory & Cognition examines how suspended goals are handled after interruption.
Education research provides a complementary reason to teach visible planning and monitoring. The Education Endowment Foundation’s 2025 metacognition guidance focuses on teaching learners to plan, monitor and evaluate. Its March 2026 practitioner article on metacognitive talk discusses teacher modelling and the gradual internalisation of these processes. The Australian Education Research Organisation’s explicit-teaching guide likewise addresses modelling and managing cognitive load.
These sources support the plausibility of explicit procedural modelling and recovery prompts. They do not establish that DONE — STATE — NEXT is a branded research intervention or that it has a measured effect size in Singapore schools. The format offered here is an original classroom design to be tested and adapted locally. That distinction matters: evidence-informed practice should not pretend a practical mnemonic carries stronger research status than it does.
What happens in Primary Mathematics when a step is repeated
A Primary 6 student works on a multi-step word problem involving total cost, discount and change. She first calculates a total, applies a ten per cent discount, and then stops to answer a sibling. When she returns, the notebook shows two amounts but no labels. She applies the discount again because the smaller number looks like a starting amount. The final answer is lower than it should be. Her multiplication is flawless; the state of the process is not.
The corrective teaching is to attach a role to every intermediate value: original total, discount, after discount, change. These labels let the learner determine what has already happened. A parent can ask, “What does this number represent at this point?” rather than immediately reteach percentage calculations. If the child cannot explain the role, return to the story of the quantities. If the child can explain but repeated the operation after interruption, practise marking a verified stage.
This example also shows the limit of checklists. If the learner never understood the meaning of “after discount,” a neat checkpoint will preserve a misunderstanding. Recovery support is not a substitute for concept teaching. It ensures that the student can carry a valid concept through a changing task without losing track of the transformations already performed.
What happens in Science when the step cannot be undone
A laboratory procedure may include actions whose effects are irreversible or not directly observable afterward. A reaction begins when substances are combined. A stopwatch start cannot be reconstructed after several uncertain minutes. A specimen may be altered by heating, cooling or exposure. If a student cannot establish whether the critical step occurred, repeating it may be unsafe or scientifically invalid. The correct action is to pause, make the apparatus safe as instructed, and consult the teacher.
At home, families should use only age-appropriate safe activities and follow school guidance; this article is not permission to improvise chemical procedures. The sugar-and-water example is deliberately simple because the learning target is recording conditions. Even there, a fair comparison requires consistency in quantities, temperature, stirring and timing. A checkpoint makes those conditions inspectable. If the first cup was stirred for an unknown period, the child should not manufacture a precise result.
Teachers can teach three recovery classes: a step that can be safely verified from the current state, a reversible step that can be corrected, and a step whose uncertainty makes the trial unusable. That is scientific reasoning about the quality of evidence, not just obedience to instructions. The student learns that “I don’t know whether I started the timer” is important information, not something to conceal.
What happens in English when revision passes overlap
A Secondary 1 student has written a composition and is told to revise it for content, organisation and language. She begins correcting spelling and punctuation because those are easy to notice. Then she remembers the teacher asked for stronger evidence in paragraph two. She rewrites the paragraph, accidentally removing a useful linking sentence. Later she checks punctuation again and leaves paragraph four’s argument untouched. Her worksheet says all three passes were completed, but the text tells another story.
A better re-entry structure distinguishes the purpose and extent of each pass. During the first pass, review claim and evidence; place a small tick beside paragraphs that have been checked. During the second, review progression and transitions. During the final pass, repair sentence-level accuracy. If interrupted halfway, mark the last paragraph genuinely reviewed for that purpose. On return, reread that paragraph to re-establish the criterion, then continue.
Students should not turn composition into a rigid mechanical assembly line. Writing is recursive: a new idea may require revision of earlier sentences. The checkpoint’s job is to preserve visibility, not forbid creative return. It lets the writer know when a section was last evaluated and for what purpose, so deliberate revision does not become accidental duplication.
A parent can help without becoming the process manager
It is tempting to stand beside a child and call out each next step: “Now do this. Have you done that? Remember the timer.” This may produce a completed task in the short term, but it can leave the parent carrying the entire state of the procedure. When the adult walks away, the child cannot re-enter independently. The more useful role is to teach the learner to build and inspect a local state marker.
Ask three questions: “Which step can you prove is done?” “What is true of the work right now?” “What is the next action that follows from that state?” Resist answering them immediately. A child may need a model the first time. Over repeated practice, reduce prompts: point to the checkpoint, then ask a single question, then invite the student to inspect it independently. That is the difference between support that builds capability and support that performs the work on the student’s behalf.
Families also need proportion. A ten-minute arithmetic page need not acquire a project-management system. Choose one recurring failure—repeating an operation after a break, losing place during a composition edit, forgetting which science condition was measured—and fix that. The child should experience the tool as relief from uncertainty, not as another task to satisfy an adult.
A classroom protocol that is short enough to survive reality
During a procedure-heavy lesson, give pupils a small working area with three columns: verified step, current state, next action. Demonstrate once using a real subject example. Then remove the scaffold for easy routines and reintroduce it when the task becomes longer or riskier. Students can work in pairs to compare whether a proposed next step genuinely follows from the recorded state, but each learner should finally be able to explain the progression individually.
A teacher can create one deliberate interruption in a low-stakes practice exercise. Students solve two steps, put the page aside briefly for an unrelated short activity, and then return. Some will resume correctly without a marker; others will need to inspect their working. Compare the quality of re-entry rather than the speed alone. Ask what made it easy to recognise the last valid step. This prompts learners to articulate procedural cues from the subject, not memorise a generic acronym.
Avoid turning this into a competition about whose checklist is prettiest. A useful state record may be ugly and terse. If it reliably prevents duplicate operations or lost measurements, it has done its job. A beautifully formatted log that is never consulted has not. The intended outcome is independence in completing and checking subject tasks, not another artefact to grade.
The hidden difficulty of instructions that cannot be observed directly
Some instructions use verbs that describe mental actions: compare, analyse, decide, evaluate, infer. A child cannot tick “understood the passage” in the same sense as “wrote a title.” Teachers can make these steps more observable without reducing them to rote outputs. For comparison, the checkpoint could be identified one similarity and one difference with evidence. For inference, claim written; evidence sentence located; explanation of link still needed. For evaluation, criterion chosen; two options considered; conclusion not written.
The aim is not to force all thinking into boxes. It is to identify a small external trace that lets the learner know what phase of reasoning has been completed. In higher-level tasks, that trace may be a question, a diagram or a margin note rather than a checkbox. Adults use the same strategy when they leave a comment in a draft: “Need to verify source before making this claim.” The note protects a partially completed cognitive job from being mistaken for a finished one.
As students gain expertise, they may need fewer explicit markers. A fluent algebra student can recognise valid transformations rapidly; an experienced writer can navigate an evolving argument without logging every paragraph. The scaffold should fade where it is unnecessary and return when the procedure is novel, complex or interrupted. Permanent dependence on a giant form would be a different failure.
When a checkpoint becomes counterproductive
Too much state recording can interrupt the task more often than the outside world does. If a student stops after every word to tick a box, the method has become a source of cognitive load. If a teacher requires a full paragraph after each arithmetic operation, the pupil may spend more time narrating than reasoning. The checkpoint must be proportional to the risk of losing state, and its information must be cheaper to record than to reconstruct.
There is also a danger of false certainty. A student ticks “measurement recorded” without checking the notebook. Later, everybody trusts the tick. The system has made an unverified claim look official. Teach pupils to use the word “verified” accurately: a result is checked against observable evidence. For irreversible or high-stakes actions, a checkpoint without verification may be worse than no checkpoint because it encourages unearned confidence.
Some learning tasks benefit from flexible exploration rather than fixed procedure. Brainstorming ideas for a story does not always need a linear stage record. A difficult Mathematics investigation may require trying and abandoning approaches. In those cases, a checkpoint can preserve the current conjecture and unresolved question without pretending there is only one correct next move. The point is to make state recoverable, not to eliminate discovery.
Five common failure modes and their repairs
The repeated-action error occurs when a student applies the same operation twice after a pause. Repair by labelling the current output and the last verified transformation. The skipped-state error occurs when an unchecked transition is assumed complete. Repair by attaching a visible result to each major step. The context-loss error occurs when a child remembers a number but not what it represents. Repair by recording units, variable names and condition labels together.
The false-completion error occurs when a tick replaces evidence. Repair by verifying the state before signing off. The unsafe-reentry error occurs when the child cannot tell whether an irreversible experimental action happened and guesses. Repair by pausing, making the task safe and asking the supervising teacher whether to restart. These repairs are not the same even though all five can appear as “forgot the instructions.”
A sixth, subtler failure is over-support. An adult manages every transition, so the child never practises recovery. The fix is graduated responsibility: model one checkpoint, practise together, then ask the learner to use the cue independently. If the task is dangerous or governed by strict safety procedures, adult supervision and formal rules take priority over an independence exercise.
The difference between forgetting a step and misunderstanding the concept
A teacher can test the difference by interrupting the routine at a known point and asking the student to explain the reason for the next step. A child who says, “I divide by two because the coefficient of x is two and I want x alone,” demonstrates conceptual understanding. If that child later divides twice because the process state was lost, reteaching what a coefficient is may be unnecessary. Teach the re-entry move.
Conversely, a learner may follow every recorded step but be unable to justify any transformation. A neat procedure can conceal fragile understanding. For example, a child solves 2x+6=18 correctly by a memorised sequence but claims that subtracting six changes only one side. Their checkpoint did not create that misconception, nor can it repair it alone. Subject teaching must still connect each action to the mathematical relationship it preserves.
The two skills should strengthen each other. Understanding why steps work makes state easier to reconstruct. Reliable records prevent the learner from having to spend attention remembering which legitimate step came next. A strong teaching sequence alternates between mechanism and execution rather than treating the checklist as proof of understanding.
A two-week local improvement test
A teacher can test whether checkpoint teaching addresses a genuine class problem using comparable multi-step tasks. During the first week, identify observable errors caused by losing state: repeated operations, missed measurements, duplicated edits, or uncertain restart points. Teach a short subject-specific state marker in the second week. Compare whether students can independently explain the last verified step after a planned pause and whether the targeted errors decrease.
This is a quality-improvement exercise, not a controlled experiment. Practice itself may improve performance. Novelty or teacher attention may change behaviour. The group may have become more familiar with the subject. It would be unsound to announce a universal treatment effect from such a comparison. But if the same student becomes better at recovering the correct step after interruption, the teacher has useful local evidence and can continue refining the support.
Track a small number of meaningful indicators: proportion of correct restarts, time spent reconstructing progress, and whether the learner can justify the next action. Avoid collecting intrusive behavioural data or building a permanent monitoring system. The best intervention should become less visible as the student becomes more independent.
Student route: the ten-second restart
When returning to interrupted work, stop before acting. Read the last line or look at the apparatus as it currently stands. Find the last step you can verify. Name what the result means. Decide whether the next action is safe and logically follows. If any of these is uncertain, consult the instructions or teacher before guessing.
On a Maths page, you might say, “The last verified equation is 2x=12, so I divide both sides by two.” In English, “I reviewed evidence in paragraphs one and two; paragraph three is next.” In Science, “Cup A has been prepared; the timer has not started.” These are small sentences, but they bridge a gap between a general plan and the actual work in front of you.
Do not be embarrassed to mark a precise place to resume. Experts use bookmarks, lab notebooks, code comments, checklists and logs for similar reasons. The point is not that competent people never forget; it is that reliable work does not depend entirely on having an uninterrupted memory of every step.
Parent route: an evening without constant prompting
Choose one task that frequently breaks after interruptions. Explain what a completed step looks like. Ask your child to write a one-line checkpoint before a planned short break. When the child returns, do not announce the next step. Ask them to use their own marker and check it against the work. If the marker is ambiguous, improve it together: “What would your future self need to know?”
Praise a verified recovery, not the decorative appearance of a checklist. If your child correctly says, “I’m not sure if that measurement happened; I should redo the trial,” that is evidence of better judgement even if it takes extra time. Scientific honesty and safe recovery are more important than forcing every activity to look uninterrupted.
If the child struggles across tasks, speak with the teacher about the specific pattern and what helps. Do not diagnose a memory or attention condition from one interrupted homework session. Fatigue, unfamiliar content, device design and household interruption can all matter. The aim is to change the visible task, not assign an identity to the learner.
Teacher route: model the invisible decision
While demonstrating a multi-step process, periodically explain why you know a step has finished. Point to the result, not just the instruction number. Ask pupils to name the evidence of completion before proceeding. Then deliberately stage a safe interruption in practice and ask them to resume from a verified state. This takes the procedure beyond “copy what I did” into “understand how to continue when circumstances change.”
Use subject-appropriate language. In algebra, state and verify equality. In Science, state and verify conditions and measurements. In writing, identify the revision criterion and the last paragraph reviewed. In digital work, distinguish typing from saving, and saving from submitting. The family resemblance across subjects is useful, but each discipline must provide the actual meaning of its checkpoints.
Finally, let support fade. A child who can resume reliably with a short margin note should not be forced to fill a three-column record forever. The purpose of a scaffold is to make a capability available until the learner can perform it independently. Local evidence—not the adult’s affection for the tool—should determine when to simplify it.
Frequently asked questions
Why does my child forget steps even after reading the instructions twice?
Reading a list does not automatically preserve knowledge of which actions are complete in the current attempt. Check whether the difficulty occurs at the same conceptual step or mainly after interruptions. The intervention depends on which mechanism is observed.
Is this just a working-memory problem?
Interruptions can increase working-memory and attention demands, but one homework incident does not diagnose an individual cognitive weakness. Task novelty, unclear instructions, missing labels and environmental interruptions may contribute.
Wouldn’t a normal checklist solve everything?
A checklist can help, but a tick without an observable result may create false confidence. A better checkpoint records the last verified state and next action, especially where a step must not be repeated accidentally.
How many checkpoints should a pupil write?
Only as many as the task needs. Use them around important transitions or predictable interruptions. Too many checkpoints consume attention that should go to the subject itself.
What if the student skips a step deliberately?
Ask why. Some procedures allow flexible order, others depend on strict sequencing. The learner should understand which steps are reversible, optional, dependent or safety-critical rather than treat every list as interchangeable.
Can a student restart a science experiment halfway through?
Only when the procedure and supervising teacher permit it. If an irreversible action or timing is uncertain, guessing is not appropriate. The safest and most scientifically valid route may be to stop and repeat the whole trial under supervision.
What is the difference between re-entry and understanding instructions?
Understanding instructions concerns what each step requires. Re-entry concerns the current state after some steps have been performed. A student can understand the list yet be unable to identify where to continue.
Does this approach apply to essay writing?
Yes, with flexibility. A writer can mark the last paragraph checked for argument, evidence or sentence accuracy. Writing is not always linear, so the checkpoint should preserve work status without forbidding creative revision.
Could technology solve the problem automatically?
Some apps provide version history, autosave indicators and progress markers. These help only if the learner understands their meaning and verifies the current state. A green Saved label does not prove that an assignment was submitted or that a calculation is valid.
Is the checkpoint formula scientifically proven?
It is a practical format informed by research on interruption, monitoring and explicit teaching. No source cited here tests this exact three-part phrase as a universal school intervention. Treat it as a design to adapt and evaluate, not as a guaranteed improvement.
Returning to the table
The Clementi student returns to the two cups and her half-completed science log. Instead of trying to reconstruct the entire event from a foggy memory, she checks what is verifiable. One cup has a label and recorded quantities. The timing for the other is uncertain. She tells the supervising adult, decides not to use the uncertain result, and prepares a fresh trial safely. The experiment takes a little longer, but its evidence becomes better.
That is an educational success even before the final graph is drawn. The learner has learned that a procedure is a changing system with a history, not merely a line of commands. The important question after interruption is not “Can I remember what I meant to do?” It is “What state can I verify, and what action follows?” A child who learns to make that distinction can bring more independence, accuracy and intellectual honesty to Mathematics, Science, writing and the many real-world tasks that rarely proceed without interruption.
Sources, qualifications and next routes
- Education Endowment Foundation: Metacognition and Self-Regulated Learning guidance — updated research-informed teaching recommendations, 2025.
- EEF: Promoting metacognitive talk in the classroom — 4 March 2026 article about modelling, planning, monitoring and evaluation.
- EEF: Metacognition and self-regulation toolkit evidence — evidence synthesis last reviewed May 2025.
- Australian Education Research Organisation: Teach explicitly — modelling and cognitive load, updated 31 March 2025.
- Scientific Reports research on anticipating task interruptions — adult laboratory research published 2024; relevant mechanism but not direct school intervention evidence.
- Memory & Cognition: Resumption costs in task-interruption contexts — research on returning to suspended task goals.
- Institute of Education Sciences: Organizing Instruction and Study to Improve Student Learning — learning and memory considerations in instructional design.
Return to How X Works for the wider education mechanism library. For a different question about preserving evidence of an erroneous attempt, see why crossed-out working matters. This page owns the point at which a learner must resume a changed procedure from a verified state.