How Successive Relearning Works | Retrieval That Returns Until It Survives Time

Quick Read: The Core Mechanism

Successive relearning can be reduced to a cycle:

Retrieve → check → correct → retrieve successfully → leave → return later → retrieve again → repeat.

Katherine Rawson and John Dunlosky define successive relearning as practising a task until it is performed correctly and then practising it again until it is performed correctly during later spaced sessions. Their 2022 review in Current Directions in Psychological Science describes it as a potent but underused route for obtaining and maintaining knowledge. Their earlier work and classroom-facing review connect the method to practice testing, feedback, criterion learning and spacing.

Notice what that definition does not say. It does not say “read the notes several times.” It does not say “do one quiz.” It does not say “look at a flashcard every day.” The important structure is criterion plus return. The learner must actually produce the knowledge, receive enough information to repair failure, and then prove later that the route still exists.

Why One Correct Recall Is Not the Same as Knowing

Imagine a student named Maya learning the word mitigate. On Monday, she studies the meaning: to make something less severe, harmful or painful. Two minutes later, her tutor asks for the meaning. Maya answers correctly.

What has been demonstrated?

Something useful—but limited. Maya has shown that the representation was available shortly after exposure. The cues are still warm. The page, the explanation and the recent conversation are all nearby in memory. She has not yet shown that the word can survive a different day, a different paragraph, a different emotional state, a different subject, or an examination paper in which nobody tells her that mitigate is the word she needs.

That gap between immediate performance and later capability is one of the central traps in education. A student can look excellent while supports are still present. A class can feel successful because everyone can answer at the end of the lesson. A worksheet can produce ninety percent accuracy because the examples are blocked together and the relevant procedure remains obvious.

Then three weeks pass.

The same learner looks as though nothing was taught.

Often something was taught. The problem is that the learning route was not required to survive repeated returns. Successive relearning turns survival into part of the learning objective rather than treating forgetting as an unfortunate event discovered at the end.

The Difference Between Repetition and Relearning

Repetition is not automatically useless. Seeing, hearing and using information more than once can matter. But repetition can happen without genuine reconstruction.

A student can reread a model answer four times and become increasingly familiar with it. Familiarity lowers friction. Sentences begin to feel predictable. The page becomes easier to move through. That subjective ease can be mistaken for mastery.

Relearning asks for a stronger event. The knowledge has become less immediately accessible, so the learner must recover it. If recovery fails, the learner repairs the route and tries again. If recovery succeeds, the learner has evidence that the route can function after some degree of forgetting.

This is why a good relearning session may feel worse than a rereading session. Rereading exposes the learner to fluent information. Relearning exposes the learner to the current state of memory. One produces a smooth experience; the other produces evidence.

For the larger distinction between activity and durable capability, begin with the Study & Learning Methods Hub and How Studying Works — Studying Is Not Learning.

Criterion Changes the Game

Many study routines are time-based. Revise vocabulary for twenty minutes. Do mathematics for forty-five minutes. Read science notes for half an hour. Time matters because human life has limits, but time alone cannot tell us whether the target was achieved.

Successive relearning introduces a criterion. A criterion might be one correct retrieval, two correct recalls in the same session, a complete explanation without notes, or accurate execution of a procedure with no critical error. The exact criterion depends on the task, but the idea is stable: the session ends because performance has reached a defined state, not merely because the clock moved.

This makes the learner’s time variable. Easy items may need one attempt. Difficult items may need several. That is educationally important because knowledge is not equally weak. A fixed worksheet gives every item the same number of appearances. Criterion learning allocates additional work where the system proves that it needs it.

The danger is overtraining a tiny item while neglecting the larger skill. Criterion should therefore match the actual job. For a multiplication fact, quick accurate recall may be appropriate. For a concept such as opportunity cost, the criterion should not be “say the definition.” It should include explaining the idea, recognising it in a new situation and distinguishing it from a nearby concept.

Spacing Creates the Test That Matters

If retrieval happens again immediately, the learner may still be operating on short-lived accessibility. Spacing inserts time, and time changes the problem.

During the gap, other classes happen. Other thoughts arrive. Sleep occurs. Similar concepts compete. Context changes. The memory route is no longer surrounded by the same cues. When the learner returns, the act of retrieval becomes a more realistic test of whether the knowledge can be found independently.

Spacing is therefore not simply “waiting because forgetting is bad.” Some forgetting is useful because it makes the next recovery meaningful. If the answer is fully active, there is little retrieval problem to solve. If the answer is completely gone, the learner may need substantial restudy. Good spacing places the return in a zone where retrieval requires effort but remains recoverable with feedback.

This is one reason there is no universal perfect interval. The useful gap depends on prior knowledge, material difficulty, desired retention period, the learner’s age, the number of competing topics and the quality of each retrieval event. The principle is more durable than any single calendar rule: return after enough time for memory to be tested, but not so late that every session becomes total reconstruction from zero.

A Four-Session Example

Suppose a Secondary student is learning eight causes and consequences related to a history topic.

Session 1: Build and Retrieve

The student first learns the material well enough to understand it. Then the notes close. Each idea must be retrieved from a question or cue. Wrong or incomplete answers are corrected. The student continues until each target meets the chosen criterion.

Session 2: Relearn

The same targets return after a gap, but the order changes. Some cues are phrased differently. The learner again attempts retrieval before seeing the answer. Items that fail receive feedback and another attempt until they return to criterion.

Session 3: Relearn Under Variation

Now some prompts ask for explanation rather than listing. One asks for comparison. One removes the familiar heading and embeds the idea in a short source. The learner is still retrieving the same knowledge, but the cues are less identical.

Session 4: Relearn for Use

The learner retrieves the core content again, then uses it in a paragraph that requires selection and connection. At this point the system is no longer merely preserving isolated facts. It is testing whether the facts remain available when the task becomes larger.

That progression matters. A relearning system should ultimately move from item survival to usable knowledge.

Successive Relearning Is Not Just Flashcards

Flashcards are a convenient implementation because they separate cue from answer. But successive relearning is a learning architecture, not a stationery format.

You can successively relearn vocabulary meanings and usage, mathematical formulas and conditions, grammar distinctions, science definitions and causal chains, historical relationships, geographical processes, literary terminology, oral-response structures, essay plans, worked procedures, proof skeletons, foreign-language forms, music theory, technical symbols, or any knowledge that must remain retrievable over time.

The tool can be paper, a whiteboard, a question bank, an oral tutor exchange, a spreadsheet, an app, a notebook or a blank sheet. The non-negotiable part is the learning event: attempt from memory, verify, repair, meet criterion, return later.

But Some Things Should Not Be Reduced to Cards

There is a common mistake in evidence-based study advice. Once a technique is shown to be powerful, people try to make every form of learning look like the technique.

That is backwards.

Successive relearning is excellent for knowledge that benefits from durable retrieval. It is not a complete theory of writing, problem solving, design, laboratory work, collaboration or judgement. A student cannot become a strong writer merely by retrieving writing facts. A mathematician cannot solve novel problems merely by recalling formulas. A scientist cannot reason from evidence merely by memorising definitions.

Durable components make complex performance possible, but components must later be coordinated. The Micro–Meso–Macro Learning Control Tower is useful here. Successive relearning can stabilise micro knowledge. The learner must then bind that knowledge into meso routines and macro performance.

A student who can perfectly retrieve the quadratic formula but cannot recognise when a problem is quadratic has durable knowledge without sufficient routing. The formula survived. The decision system did not.

The Recognition Trap

One reason students underuse retrieval-based methods is that recognition feels persuasive.

Look at a page you have read three times. It is familiar. Headings feel known. Examples feel expected. When the answer appears, it seems obvious. The brain quietly converts “I recognise this” into “I could have produced this.”

Close the page and the illusion is exposed.

This is a measurement problem. The study method is giving the learner access to the answer while asking the learner to estimate independent access. Those are different states.

Successive relearning improves the measurement by repeatedly removing the answer. Each return is a new audit. The learner does not have to guess whether the memory is durable; the system keeps asking it to prove that it can still operate.

Why Difficulty Can Be Useful Without Worshipping Difficulty

Evidence-based learning often produces a paradox. Methods that create stronger long-term learning can feel less fluent during practice. Retrieval can feel harder than rereading. Spacing can feel worse than massing. Interleaving can make practice scores drop before transfer improves.

That does not mean “harder is always better.” Difficulty is useful only when it forces the learner to perform operations that matter and when the learner still has a realistic route to success.

A question that is impossibly obscure is not automatically educational. A delayed review so late that the learner has forgotten everything may waste time. A retrieval cue so vague that many answers could be defensible may measure confusion rather than memory.

The aim is not suffering. The aim is reconstruction.

See also How Desirable Difficulty Works.

Feedback Is the Repair Crew

Retrieval without feedback can preserve errors as efficiently as it preserves correct knowledge.

Suppose a learner repeatedly retrieves that the area of a circle is 2πr². The act of recall is effortful. It is spaced. It is repeated. It is also wrong.

Successive relearning therefore needs a trustworthy answer source. After the attempt, the learner checks. If the answer is wrong or incomplete, the correction must be noticed, understood and then produced again. Merely glancing at the correct answer is often too weak. The corrected route should itself be retrieved.

• Attempt reveals the current state.
• Feedback identifies the gap.
• Correction rebuilds the representation.
• Successful retrieval demonstrates that the repair can now run.
• Later return tests whether the repair survived.

That loop is more useful than simply counting how many questions were completed.

The Learning Ledger: Track States, Not Streaks

Apps often reward streaks because streaks are easy to display. A learner sees seven consecutive days and feels momentum. Momentum can be useful, but a streak does not tell us whether the underlying knowledge is stable.

A more meaningful relearning ledger tracks states. An item may be new, fragile, retrievable with cue, retrievable independently, stable across two intervals, stable under variation, or usable in transfer. These labels need not become bureaucratic. Their purpose is to change the question from “Did I study?” to “What state is this knowledge in now?”

That shift is powerful because learning is not binary. A concept can exist while still being unreliable. A procedure can be accurate but slow. A vocabulary word can be understood receptively but unavailable expressively. A historical relationship can be recalled only when the topic heading is supplied.

Successive relearning makes those intermediate states visible because the same target is observed through time.

Successive Relearning for Vocabulary

Vocabulary is one of the clearest applications because word knowledge has several layers. A word may be recognised in reading long before it can be retrieved in speaking or writing.

For the word reluctant, a weak card might ask only “reluctant = ?” A stronger relearning sequence changes the retrieval job across sessions. Give the meaning in your own words. Distinguish reluctant, hesitant and unwilling. Produce a sentence where reluctance is implied by behaviour. Explain what would make the word inappropriate in another sentence. Retrieve an antonym. Use the word naturally in a paragraph.

The knowledge being successively relearned is no longer a dictionary line. It is a semantic network. That matters because real language use asks for selection under context.

For the wider vocabulary architecture, use the Vocabulary Learning Hub.

Successive Relearning for Mathematics

Mathematics creates a different challenge. Some knowledge should become fast and stable: number facts, algebraic identities, formula relationships, angle properties, standard forms and definitions. But mathematics is damaged when retrieval is mistaken for understanding.

A useful mathematics relearning system separates at least three layers.

• Retrieve the object: can the student recall the formula, definition or rule accurately?
• Retrieve the conditions: when does the rule apply and what assumptions are required?
• Recognise the need: can the student identify the structure in an unfamiliar question without being told which formula to use?

The first layer is memory. The second is conceptual control. The third begins transfer. Successive relearning should support all three without collapsing them into one flashcard score.

Use the Mathematics Learning Hub when the goal moves from stable knowledge to mathematical performance.

Successive Relearning for Science

Science contains vocabulary, mechanisms, models, evidence and causal reasoning. A learner may remember every noun in a process but fail to explain the process itself.

For photosynthesis, for example, successive relearning might begin with core inputs and outputs, but later sessions should require the learner to reconstruct relationships: where the process occurs, what energy transformation is involved, what limits the rate, how the process connects to respiration, and how evidence could distinguish competing explanations.

The best retrieval prompt often has a verb. Explain. Compare. Predict. Sketch. Justify. Distinguish. Trace. Verbs force knowledge to move.

For subject-specific routes, enter the Science Learning Hub.

Successive Relearning for English

English is especially vulnerable to oversimplified study advice because much of the subject is performance rather than declarative recall. Students need vocabulary, grammar, text structures, literary concepts and response frameworks, but they also need judgement.

Successive relearning can stabilise the knowledge that supports judgement. A student can repeatedly retrieve the distinction between fact and inference, the purpose of a counterargument, the conditions for a comma splice, common rhetorical moves, or the features of a situational-writing task.

Then the retrieval must be embedded in fresh texts. The student should not only define “tone” but infer tone from a new passage. Not only list persuasive techniques but explain why a particular technique matters for a particular audience. Not only recall the structure of a paragraph but construct one under a new question.

That transition from durable knowledge to flexible performance is where the English Learning Hub becomes the next route.

The Hidden Variable: Cue Quality

Retrieval is always retrieval from something. A question, image, phrase, problem, heading or context acts as a cue. If the cue is too generous, the task may be easier than real performance. If the cue is too weak, the task may become ambiguous.

Compare three prompts: “What is the formula for speed?”; “A cyclist travels 18 km in 45 minutes. What relationship do you need?”; and “Which quantities in this unfamiliar transport problem determine the rate, and why?” All three can involve the same underlying relationship, but the cue moves from explicit retrieval to structural recognition.

A mature relearning system gradually varies the cue so that the learner does not become dependent on one familiar wording. This is also why copying a teacher’s exact question set forever is risky. The student may learn the route from the wording rather than the route from the concept.

The Hidden Variable: Response Quality

A learner can technically retrieve an answer that is too shallow for the real task. “Because friction” may be accepted in a casual quiz but fail in a science explanation that requires mechanism. “Democracy means voting” may retrieve a partial association while missing institutions, rights, accountability and rule structures.

Criterion therefore needs quality boundaries. For complex knowledge, “correct” should sometimes mean complete enough for the task, precise enough to exclude a misconception, connected to the relevant mechanism, expressed in the learner’s own words without changing the meaning, and usable in a new context.

That turns successive relearning from a memory game into an architecture for dependable knowledge.

What Should Be Relearned First?

Not everything deserves equal repetition. A learner with limited time should prioritise knowledge with high future leverage.

• foundational knowledge used repeatedly in later topics;
• items that are easy to confuse;
• knowledge that repeatedly fails under exam conditions;
• high-frequency vocabulary or concepts;
• procedural steps where one missing component breaks the whole chain;
• core representations that reduce working-memory load;
• and knowledge required for transfer into larger tasks.

This is a bottleneck principle. The most valuable retrieval target is not necessarily the most recent item. It is often the item whose absence forces the learner to stop, guess or rebuild repeatedly.

For diagnosing that problem, see the Diagnostics & Recovery Hub.

What Should Be Removed From the Relearning Queue?

A good system also knows when to stop spending attention on an item.

If an item is repeatedly retrieved with high accuracy across long intervals and varied cues, continuing to test it at the same frequency has diminishing value. The interval can expand. The item can move into a maintenance pool. Time can be reassigned to weaker or more consequential knowledge.

This prevents one of the most common flashcard failures: the learner builds an enormous collection and becomes a maintenance worker for the collection itself. The tool begins consuming the study plan.

The goal is not to keep cards alive. The goal is to keep knowledge available for meaningful use.

When the Calendar Lies

Students often build revision schedules that are aesthetically satisfying but cognitively weak. Monday is Chapter 1, Tuesday is Chapter 2, Wednesday is Chapter 3. Each box gets ticked once. The plan looks complete because the calendar has been covered.

But a calendar tracks encounters, not survival. If Chapter 1 never returns after Monday, the plan assumes that exposure is equivalent to retention. Successive relearning changes the geometry of the schedule. Old material keeps reappearing in small doses while new material is added.

This means a good revision week can look less tidy. Tuesday contains some Monday. Friday contains some Tuesday. Next week still contains a small amount of this week. The curriculum moves forward while memory loops backward.

That is not inefficiency. It is how cumulative learning protects itself from time.

The Exam Problem: Retrieval Must Survive Pressure

A student may retrieve successfully at a quiet desk and still fail during an examination. That does not invalidate the earlier learning. It reveals another layer.

Exams change cue structure, time pressure, emotional load, task switching and the cost of error. The student may know the material but fail to access it quickly enough. Or the relevant cue may be hidden inside an unfamiliar question.

Late-stage successive relearning can therefore borrow features from target performance: timed retrieval, mixed topics, reduced cueing, brief written explanations and rapid decision prompts. The aim is not to turn every early study session into a mock exam. It is to gradually close the distance between learning conditions and use conditions.

For that performance layer, use the Examinations & Assessment Hub.

Why Students Often Avoid the Method

Successive relearning asks students to encounter forgetting repeatedly. That can be psychologically inconvenient.

Rereading says: “Look how much you recognise.” Retrieval says: “Show me what remains when the page disappears.”

The second method can make a capable student feel less capable during practice. A learner who confuses performance with identity may interpret every retrieval failure as evidence that they are bad at the subject.

A teacher or parent can change the meaning of the event. Failure during retrieval is not a verdict. It is location data. It tells us where the route is thin enough to require repair before the stakes become high.

The safest place to discover forgetting is during practice, while correction is cheap.

A Practical Student Protocol

For a normal school week, successive relearning can be implemented without turning life into a laboratory.

• Select: choose a small set of important targets.
• Understand: do not memorise words you do not understand.
• Retrieve: close the source and attempt from memory.
• Verify: compare against a trustworthy answer or rubric.
• Repair: correct the missing or wrong part.
• Meet criterion: retrieve accurately enough before ending the item.
• Space: schedule a later return.
• Vary: change the cue or use the knowledge differently.
• Expand interval: strong items return less often.
• Transfer: use the knowledge inside a larger unfamiliar task.

The protocol is intentionally simple. Complexity belongs in the content, not in the bookkeeping.

A Parent Protocol: Ask for Evidence, Not Hours

Parents often receive only a time report: “I studied for two hours.” That tells us effort occurred, but not what changed.

A better conversation asks: What could you retrieve today without looking? What failed? What did you correct? What will you test again later? Which items are now stable enough to leave longer? Can you use one of them in a new question?

This turns the parent from a timekeeper into a reader of learning evidence. It also reduces the temptation to reward long study hours filled with low-yield activity.

A Teacher or Tutor Protocol: Design the Return

The teacher’s job is not merely to present a concept clearly. It is to engineer future encounters with the concept.

A lesson sequence can include brief retrieval at the start of later lessons, cumulative questions, low-stakes quizzes, oral prompts, mixed practice and delayed explanation tasks. The key is that earlier learning must re-enter the room after the class has moved on.

This changes curriculum pacing. “Covered” no longer means “finished.” A topic can be formally completed while still remaining active in the retrieval architecture.

For broader instructional routes, see How Retrieval Practice Works, How Spaced Practice Works and How Formative Assessment Works.

Successive Relearning and the First Weak Link

A repeated failure can reveal more than weak memory.

If a student repeatedly forgets the same chemistry relationship, the first weak link may be an unstable prerequisite. If a vocabulary item never survives, the learner may not understand its semantic boundary. If a mathematics rule is repeatedly misapplied, the issue may be classification rather than recall.

Successive relearning is therefore diagnostic as well as instructional. Each return produces a tiny longitudinal record. What survives? What collapses? What requires more cues? What becomes stable? What remains dependent on the same wording?

Over several sessions, patterns emerge. The learner stops being a single score and becomes a changing system.

The Difference Between Maintenance and Growth

Education has two different jobs that are often confused.

One job is growth: acquiring a new concept, method, representation or skill. The other is maintenance: keeping important knowledge usable after acquisition.

Successive relearning is unusually strong because it sits across the boundary. Early sessions contribute to acquisition; later sessions protect maintenance. This matters in cumulative subjects where forgotten foundations create future learning costs.

A student who forgets fractions does not only lose a Primary topic. Ratios, percentages, algebraic manipulation, rates and later mathematics may become more expensive. A student who loses core grammar distinctions pays the cost in every new writing task. A student who forgets scientific vocabulary spends working memory reconstructing language instead of reasoning about mechanisms.

Maintenance is not glamorous. It is infrastructure.

Relearning as Compression of Future Effort

The value of relearning is not only that the learner knows more later. It can change the cost of future learning.

When foundational knowledge remains available, later lessons can begin at a higher level. The teacher does not need to rebuild every prerequisite. The student does not have to hold basic relationships in working memory through conscious effort. Explanations can move faster because earlier concepts are already compressed into usable chunks.

This is why forgetting has compound costs. The immediate loss is one item. The second-order loss is that every future task depending on that item becomes heavier. Successive relearning is therefore a kind of preventative maintenance on the learning system.

Students often feel that returning to old material steals time from new material. Sometimes it does. But if the old material is a dependency, preserving it may be what makes the new material learnable at all.

When Successive Relearning Becomes Inefficient

No method should be protected from criticism simply because research supports it.

Successive relearning becomes inefficient when the learner memorises low-value material, the criterion is too strict for trivial facts, the queue grows faster than it can be maintained, the same cue is repeated until cue dependence masquerades as mastery, items remain isolated and never enter authentic tasks, feedback is unreliable, or the technique displaces reading, discussion, creation, problem solving and other necessary forms of learning.

The solution is not to abandon the method. It is to preserve the job boundary. Use successive relearning for durable access. Use other methods for other jobs.

Why the Method Scales From Primary School to Adulthood

The content changes, but the architecture remains recognisable.

A Primary learner can successively relearn number bonds, spelling patterns and science vocabulary. A Secondary learner can relearn algebraic identities, language structures and historical relationships. A Junior College learner can maintain definitions, equations and argument frameworks. An adult can relearn professional terminology, medical procedures, legal rules, programming syntax, language vocabulary or technical standards.

The adult advantage is not a different brain. It is often better control over selection and scheduling. Adults can choose what deserves maintenance based on actual future use.

This is why lifelong learning is not simply continuous acquisition. It is also intelligent forgetting and intelligent maintenance: deciding what may safely decay, what must stay accessible, and what should become so familiar that it no longer consumes scarce attention.

Successive Relearning as a Learning Supply Chain

There is another way to see the method. Knowledge has to move from first exposure to future use. Along that route it can be lost, distorted, delayed or made inaccessible. Each relearning session is a checkpoint in the supply chain.

The first session manufactures a usable representation. Feedback performs quality control. Spacing creates storage time. Later retrieval checks inventory. Variation tests whether the item can be delivered under different conditions. Transfer proves that the knowledge reaches the final customer: the real task.

This metaphor is imperfect, but it reveals why one successful classroom performance is insufficient. A product that worked at the factory door is not yet proof that the supply chain can deliver it intact weeks later.

From Memory to Intelligence

Durable memory is not intelligence by itself. But intelligence becomes expensive when useful knowledge must be reconstructed from zero every time.

Stable knowledge frees attention for higher-order decisions. A reader who instantly recognises common words can think about the argument. A mathematician who retrieves foundational relationships can focus on structure. A musician who has automatised scales can attend to expression. A scientist who retains core principles can reason about the unfamiliar case.

Successive relearning therefore belongs to a larger architecture of capability. It protects the components that future thinking will need. It reduces the chance that the next difficult task becomes a simultaneous test of reasoning, recall and reconstruction.

The Final Upgrade: Relearn Connections, Not Only Nodes

Facts are nodes. Education becomes powerful when the edges between them are also durable.

A student can remember “evaporation,” “condensation” and “precipitation” separately yet still hold a weak model of the water cycle. Another can know “claim,” “evidence” and “reasoning” but fail to connect them in an argument. A mathematics learner can recall “gradient” and “rate of change” as vocabulary without seeing that they describe the same structural relationship in different contexts.

Advanced successive relearning therefore asks relational questions: How are these two ideas connected? What changes if one variable increases? Why does this concept belong with that one? What would be a counterexample? Which representation preserves the same relationship? What earlier knowledge does this depend on?

Now the learner is not only keeping information alive. The learner is keeping the map alive.

The Quiet Power of Coming Back

Most educational narratives celebrate beginnings: the new topic, the new book, the new course, the new explanation, the moment something finally makes sense.

Successive relearning celebrates something less cinematic: return.

You come back on Wednesday. You come back next week. You come back after other subjects have filled the day. You ask the same knowledge to stand up again, perhaps from a different cue. Sometimes it does. Sometimes it does not. When it fails, you repair it while the stakes are low.

Over time, the return changes. What once required a page of explanation can be reconstructed from a small cue. What once felt fragile begins to behave like infrastructure. The learner spends less time asking, “Have I seen this?” and more time proving, “Can I use this now?”

That is a better definition of durable learning.

Research and Further Reading

• Rawson & Dunlosky — Successive Relearning: An Underexplored but Potent Technique for Obtaining and Maintaining Knowledge
• Rawson, Vaughn, Walsh & Dunlosky — Investigating and explaining the effects of successive relearning on long-term retention
• Dunlosky & Rawson — Practice Tests, Spaced Practice, and Successive Relearning
• Carpenter, Pan & Butler — The science of effective learning with spacing and retrieval practice
• Study & Learning Methods Hub

---

eduKateSG Learning Node Series · 0001 of the continuing series. Continue through the Study & Learning Methods Hub and the wider eduKateSG Learning Hubs.