How Studying Works | Working-Memory Strategy Emergence — Why Better Performance Can Come From a Better Method, Not a Bigger Memory

HSW-0197 · How Studying Works

A student is asked to remember six instructions.

On Monday, the student tries to hold every word exactly as heard and loses the middle.

On Thursday, the student quietly groups the instructions into three pairs, turns two of the pairs into little images, and remembers almost all of them.

Did the student suddenly acquire a larger working-memory capacity?

Not necessarily.

The student may have learned to operate the same limited workspace more intelligently.

Working-memory strategy emergence is the development or discovery of deliberate ways to encode, maintain, transform, prioritise and retrieve information held briefly in mind.

That distinction matters because working-memory performance is never only about a fixed “size.” It also reflects what the learner does with the material, which strategies the task permits, whether those strategies can be executed accurately, and whether the learner knows when to switch methods.

This article owns the narrow problem of strategy use inside working-memory tasks. It does not replace the broader canonical owners for working memory, retrieval working-memory load, or generic study strategy selection.

The 50-Second Read

• Working-memory scores reflect strategy as well as capacity. Rehearsal, grouping, visualisation, reordering and prioritisation can change performance.
• Children use strategies earlier and more often than a simple “capacity” story suggests. A 2026 study of 7–10-year-olds found strategy use was common across several working-memory tasks.
• Strategies are task-dependent. A method that helps backward recall may not be the best method for following instructions or n-back monitoring.
• More sophisticated is not automatically better. A cognitively expensive strategy can fail if the learner cannot execute it reliably.
• Training a task can improve the task without proving broad capacity growth. Strategy learning is one reason near-task gains must be interpreted carefully.
• Good teaching makes strategies visible, then tests whether the learner can select them independently.
• The practical loop: task demand → current strategy → error pattern → alternative strategy → independent retest.

1. Working Memory Is a Workspace, but the Learner Still Needs a Work Method

Working memory is often described as a limited mental workspace: the temporary system that lets us hold information active while using it.

That description is useful, but incomplete for studying.

Two learners with similar underlying limits can perform differently because one learner has a better procedure for handling the contents of the workspace.

Think about remembering the spoken sequence:

blue triangle — red circle — yellow square — green star — black moon — white arrow

One learner may repeat every phrase. Another may compress the six items into three visual pairs. Another may prioritise only shape, then reconstruct colour from a separate association. Another may invent a path through a familiar room.

The memory demand is not identical anymore because the learner changed the representation.

2. What the 2026 Child Study Found

A 2026 open-access study in Memory & Cognition interviewed 63 children aged 7–10 after they completed five working-memory tasks, including backward span, following-instructions tasks and an n-back task. Strategy use was common: most children reported using one or more strategies across the tasks, and strategy use was generally associated with stronger task performance. The authors also found that children often relied on one dominant strategy and that manipulation strategies were frequently reported. See Hrysanidis and colleagues, 2026.

The study is especially useful because it does not reduce strategy to one technique. Children reported rehearsal, grouping, pattern finding, semantics, visualisation, imagery, reversing at input, prioritisation, updating and other task-specific approaches.

But the evidence boundary matters. This was an exploratory analysis of self-reported strategy use in a subset of children from a training trial. Associations between strategy use and performance do not prove that a particular strategy caused the higher score, and the parent trial did not show broad intervention benefits up to six months.

That combination is educationally important: strategy use can matter without supporting the claim that generic working-memory training enlarges a general-purpose mental capacity.

3. Adults Show the Same Basic Problem: Different Tasks Invite Different Strategies

A 2024 Scientific Reports study examined self-reported strategies across several adult working-memory tasks. Strategy use varied substantially with the task and stimulus characteristics, and more sophisticated manipulation strategies were associated with better performance in some paradigms. See Ritakallio and colleagues, 2024.

The lesson is not “use the most complicated mnemonic available.”

The lesson is that performance partly reflects a match:

strategy × material × task demand × learner capability.

4. Maintenance Strategies and Manipulation Strategies

A useful distinction separates strategies that mainly maintain information from strategies that transform it.

Strategy family	Typical move	Possible use	Risk
Maintenance	Repeat, verbalise, refresh	Short lists, exact wording	Consumes time and may collapse with long sequences
Grouping	Combine items into chunks	Structured lists, categories	Bad groups can distort order
Visualisation	Turn items into an image or spatial arrangement	Concrete material	Can add irrelevant detail
Reordering	Reverse or transform sequence	Backward span, ordered operations	Transformation itself uses capacity
Prioritisation	Protect high-value elements first	Complex instructions	Lower-priority details may be lost
Updating	Continuously replace older items	Monitoring changing information	High control demand

No row is universally best.

5. Strategy Emergence Can Look Like Capacity Growth

Suppose a learner practises a memory task for two weeks and improves from four items to seven.

Several mechanisms could produce that improvement:

• better familiarity with task rules;
• faster encoding;
• more efficient rehearsal;
• chunking;
• anticipating the response format;
• reducing avoidable errors;
• discovering a task-specific shortcut;
• genuine changes in underlying cognitive efficiency.

The score alone cannot tell us which mechanism changed.

This is why transfer matters. If improvement survives changed materials, changed response formats and unfamiliar tasks, the evidence for a broader capability gain becomes stronger. If improvement collapses when the task changes, a task-specific strategy may have carried much of the gain.

6. The Utilisation Deficiency Problem

Children can sometimes know a strategy and even use it without gaining from it.

This is often called a utilisation deficiency: the strategy is available, but its execution cost is high enough that the learner does not yet receive a net benefit.

Imagine a child learning to group a six-item sequence into pairs. The grouping procedure itself may initially consume attention. The learner now has two jobs:

• remember the original material;
• operate the new grouping system.

Until the strategy becomes smoother, performance may stay flat or even worsen.

This matters for teaching. A strategy should not be abandoned after one awkward attempt if the learner is still learning how to execute the strategy itself.

7. Mathematics: Hold Structure, Not Every Symbol

Consider mental algebra.

A learner solving 3(x + 4) − 2x can try to hold every intermediate expression exactly. Or the learner can compress the task into structural moves:

1. expand;
2. collect x-terms;
3. collect constants.

The second learner has created a strategy that reduces representation burden.

As algebraic schemas strengthen, several operations become a single meaningful chunk. What once occupied multiple working-memory slots can become one familiar structure.

8. English: Use a Structure for a Long Sentence

A complex sentence can overload a reader who attempts to retain every word equally.

A stronger strategy is to hold roles:

• main claim;
• qualification;
• evidence;
• contrast;
• conclusion.

The words remain important, but the learner has reorganised the temporary representation around meaning.

9. Science: Instructions Are Easier When Converted Into a Procedure Map

Multi-step science instructions often fail because learners try to retain a spoken stream.

A strategy can convert the stream into:

prepare → change one variable → measure → record → compare.

The learner is no longer remembering five unrelated sentences. The learner is maintaining a procedural skeleton and attaching details to it.

10. Strategy Choice Is a Metacognitive Skill

Knowing several strategies is not enough.

The learner must also know:

• what kind of task is present;
• which representation is fragile;
• which strategy is likely to help;
• whether the strategy is working;
• when to switch.

This is where working-memory strategy use connects with metacognition.

EEF’s current guidance on metacognition and self-regulation emphasises explicit teaching of planning, monitoring and evaluating strategies in the context of subject learning rather than assuming these skills appear automatically. See EEF Metacognition and Self-Regulated Learning.

11. The Strategy Interview

After a task, ask:

• What did you do to keep the information in mind?
• Did you repeat it, group it, picture it, reorder it or focus on part of it?
• At what point did the method stop working?
• What changed when the list became longer?
• Would you use the same method for a different kind of task?

This is more informative than asking only, “Was it hard?”

The question exposes the learner’s operating method.

12. The Strategy-Switch Diagnostic

If performance stalls, change one dimension at a time.

1. Run the task with the learner’s normal strategy.
2. Record error type: omission, order error, intrusion, confusion, late collapse.
3. Introduce one alternative strategy.
4. Practise the strategy on an easier version first.
5. Return to the original difficulty.
6. Change the material while keeping the strategy.
7. Check whether the gain survives after a delay.

This distinguishes “I cannot hold it” from “I am holding it badly.”

13. Center-to-Edge: Build Strategy From the Core Demand

1. Center: identify what must remain available.
2. First ring: identify the dominant loss—order, identity, relation or instruction.
3. Second ring: choose one strategy that directly protects that information.
4. Third ring: practise the strategy until its operating cost falls.
5. Edge: test the strategy under changed content, speed and response conditions.

The goal is not to collect mnemonics. It is to build adaptive control over a limited workspace.

14. The School Route: Make Invisible Strategies Discussable

Teachers see answers. They do not automatically see the temporary operations that produced the answers.

Two students can give the same correct response while using very different internal procedures.

Occasional strategy talk can reveal:

• which students rely only on repetition;
• which students spontaneously group information;
• which students overcomplicate simple tasks;
• which students have a good method but execute it unreliably;
• which students need the strategy modelled explicitly.

That is actionable teaching information.

15. The Systems Route: Performance Is Capacity Plus Control Policy

A computer with fixed memory can still perform very differently under different memory-management policies.

The analogy is imperfect, but useful.

Human working-memory performance depends not only on limits but on allocation:

• what is kept active;
• what is transformed;
• what is allowed to fade;
• what gets grouped;
• what becomes the retrieval cue.

Strategy is a control policy over limited cognitive resources.

16. The Financial Route: Strategy Reduces Operating Cost

Think of each item held separately as a small ongoing cost.

Chunking, pattern recognition and procedural structure can reduce the number of independent units that must be actively managed.

But strategy has an acquisition cost. A learner must invest effort to learn and automate the method.

The useful question is not “Is this strategy clever?” but:

Does the strategy reduce net cognitive cost after its learning cost is included?

17. The Learning Route: Teach One Strategy Against One Failure

Do not give a learner ten memory tricks at once.

Attach one method to one recurring failure.

• Order errors → structured sequencing.
• Middle-item loss → chunking or grouping.
• Instruction collapse → convert to action units.
• Visual confusion → spatial patterning.
• Rapid replacement demands → updating strategy.

Once the learner can use the method, create contrast cases requiring a different strategy.

18. The Education Route: Do Not Mistake Task Gains for General Cognitive Growth

This is the most important evidence boundary.

If a learner improves on a practised working-memory task, ask whether the improvement transfers.

Can the learner:

• follow longer classroom instructions;
• hold intermediate mathematics results;
• read more complex sentences;
• solve an unfamiliar problem with less external support?

If not, the training may have improved a task-specific method rather than a broad underlying capacity.

19. The Training Route: Strategy Fading

Use a simple progression:

1. Model the strategy aloud.
2. Use it together on easy material.
3. Let the learner perform while naming each step.
4. Remove the verbal prompt.
5. Change the material.
6. Offer two plausible strategies and require a choice.
7. Remove strategy labels entirely.
8. Retest after a delay.

The final target is not strategy compliance. It is independent strategy selection.

20. The Improvement Route: Measure Strategy Flexibility

Track more than accuracy.

• strategy chosen;
• time to start;
• error type;
• whether the strategy fit the task;
• whether the learner switched after failure;
• whether the learner can explain why the strategy fits.

A learner who knows when not to use a strategy is becoming more capable than a learner who performs one memorised method perfectly.

21. The World Route: Experts Look Larger Because Their Representations Are Better Organised

An experienced chess player, programmer, clinician or engineer may appear to hold an impossible amount in mind.

Part of the explanation is that expertise changes representation. Familiar patterns become chunks. Irrelevant detail is ignored. High-value relations are prioritised. Temporary information is attached to long-term schemas.

Expertise does not remove cognitive limits. It changes what counts as one meaningful unit.

22. What Not to Do

• Do not interpret every working-memory score as pure capacity.
• Do not assume a strategy causes better performance because higher performers report using it.
• Do not teach a complex strategy without first lowering the task difficulty enough to learn the strategy.
• Do not force visualisation on material that is easier to maintain verbally.
• Do not confuse near-task improvement with far transfer.
• Do not treat children who use simple strategies as incapable; strategy repertoires develop.
• Do not keep scaffolding a strategy after the learner can select it independently.

23. Evidence Boundary

The current evidence supports the claim that strategy use is common, variable and associated with performance in working-memory tasks. It also shows that strategy profiles depend on task characteristics and age.

It does not support a universal ranking of strategies, a claim that all observed score improvement is strategic, or a claim that teaching one mnemonic broadly increases intelligence or working-memory capacity.

24. Return: The Workspace Matters, but So Does the Work Method

When a learner says, “I cannot hold all of this in my head,” there are at least two questions.

How much information is the task asking the learner to coordinate?

And what is the learner doing with that information while trying to hold it?

Working memory is limited. Strategy changes how those limits are used. Diagnose the failure, teach one better operation, practise until its cost falls, then test whether the learner can choose and transfer the method independently.

Continue through Working Memory, Retrieval Working-Memory Load, Learning Discrimination, the How Studying Works Numbered Series Reading Index and the How X Works Hub.