When Anxiety Changes the Learning Conditions

The most supportive tool is not necessarily the calmest or simplest one. It is the one that responds to what is difficult while keeping the intended learning within reach.

When anxiety appears during learning, it is tempting to search for an anxiety-friendly tool.

Something shorter.

More visual.

Less demanding.

A set of flashcards instead of a page of notes. A video instead of written instructions. A checklist for every step. Encouraging messages beside the difficult questions.

Any of these changes might help.

But anxiety does not create the same difficulty in every learning task. Nor is a tool supportive simply because the learner feels less anxious while using it.

A checklist may free attention that was being consumed by remembering several instructions at once. It may also do little for a learner who understands the instructions but is frightened that somebody will judge the result.

A worked example may reveal the steps in an unfamiliar process. It may become less useful when the learner needs to decide which process applies without being shown.

A prompt may make it possible to begin. If the prompt continues to supply the decision the learner is meant to develop, completing the task may conceal the part that remains out of reach.

The useful question is not:

Which tools are good for anxious learners?

It is:

What is anxiety changing here—and what must the support make possible?

Anxiety does not identify the learning problem

Anxiety can affect attention and performance. Thoughts about what may go wrong can compete with the task for limited mental resources, making it harder to direct and sustain attention efficiently. Research has also found relationships between anxiety, working memory and academic performance, although these relationships are not uniform across learners or situations. Anxiety does not simply switch off the thinking parts of the brain, and it does not affect every person or every kind of performance in the same way. (Eysenck et al., 2007; Owens et al., 2008; Owens et al., 2014)

This complexity matters.

Two learners may both say:

I can’t do this.

One may not understand what the question is asking.

One may understand it but be unable to hold all the steps in mind.

One may know how to answer and be afraid of getting it wrong.

One may be able to complete the task privately but lose access to familiar thinking when somebody is watching.

One may have encountered a small variation and no longer recognise which knowledge applies.

One may be exhausted before the task begins.

Their anxiety may be real in every case.

But anxiety is not yet a sufficiently precise description of what the learning support needs to do.

It tells us something important about the learner’s experience. It does not tell us whether they need an explanation, an example, a memory aid, greater predictability, less public evaluation, supported practice with uncertainty—or relief from a demand that is currently unreasonable.

These different difficulties can look remarkably similar from the outside.

The learner hesitates. Avoids beginning. Seeks repeated reassurance. Moves through the work quickly. Says they do not know. Becomes upset after an error. Stops when something changes.

If we respond only to the visible anxiety, we may make the experience calmer without discovering what has made the learning difficult.

Begin with what the learner is being asked to do

Before choosing a tool, it helps to identify the intended learning.

Does the learner need to:

• recognise the correct answer?
• recall something without being shown?
• explain an idea?
• follow a procedure?
• select the appropriate procedure?
• use knowledge in a familiar situation?
• respond when the situation changes?
• notice and correct an error?
• make a judgement when there is no single prescribed answer?

These are not interchangeable capabilities.

A learner may recognise the right formula when it appears on a card but be unable to recall it independently. They may recall the formula but not recognise when it applies. They may follow a worked solution yet become lost when one element of the problem changes.

The tool has not necessarily failed. It may have supported one part of the learning while leaving another untouched.

Research-based recommendations reflect these distinctions. Worked examples can reduce unnecessary demands while a learner is becoming familiar with a process. Active retrieval can strengthen access to information. Connecting concrete and abstract representations can help learners understand what formal ideas refer to. Explanatory questions can support understanding beyond answer production. Each approach acts on a different part of learning; none is a general solution for anxiety. (Institute of Education Sciences practice guide)

This gives us two questions rather than one:

What is anxiety making harder in this moment?

And:

What capability is this task meant to develop?

A useful support must respond to the first without quietly replacing the second.

Notice where anxiety enters the task

Anxiety may not sit across the whole learning experience. It may become most visible at a particular point.

Before beginning

The learner may not know how large the task is, what counts as finished or which action comes first.

A defined starting point, visible boundaries or a short plan may reduce the work involved in entering the task.

But difficulty beginning can also come from anticipated failure rather than unclear organisation. Breaking the task into smaller pieces will not necessarily resolve the belief that the first answer must be right.

When the instructions must be interpreted

The language may be ambiguous, unfamiliar or dense. The learner may be trying to work out both what the question means and how to answer it.

Clarifying the task is not lowering the expectation. It removes a source of difficulty that may have little to do with the intended learning.

However, if interpreting this kind of language is itself part of the destination, repeatedly translating every instruction may prevent that capability from becoming visible.

When several elements must be managed together

A learner may understand each step separately but lose track when they must coordinate them.

A task planner, partially completed example, diagram or written sequence may make intermediate information available outside working memory. This can allow attention to remain with the part that requires thought.

The appropriate support depends on what ultimately needs to be available without assistance. In many real situations, using a checklist or reference guide is part of competent performance. In others, important information must be recalled quickly or used when the familiar aid is unavailable.

When an answer must be retrieved

Seeing information and bringing it to mind are different experiences.

A learner may feel comfortable while rereading notes and suddenly conclude that their knowledge has disappeared when asked to retrieve it. The difficulty of retrieval can feel like evidence that learning has failed, even though the attempt itself can help strengthen later access to the information.

But retrieval practice does not have to begin as a high-stakes test. It can begin with cues, choices, partial prompts or permission to check—and change as the learner develops a more reliable route to the answer.

When the work becomes visible to somebody else

The learning may be accessible in private and far harder when the learner expects evaluation.

Attention that was available for the task may be pulled toward signs of judgement:

Are they watching how long I take?

Will they think this answer is ridiculous?

What will getting it wrong say about me?

More explanation will not necessarily solve a difficulty created by exposure. Early practice may need to be separated from public performance. Feedback may need to stay close to the particular decision, sentence or step rather than expanding into a judgement about the learner.

After an error

An error can supply useful information, but only if the learner can remain in contact with it long enough to examine what happened.

Research on learning from errors suggests that emotions, beliefs, the surrounding error climate and the form of feedback can all influence whether learners examine, correct or avoid errors. Errors are not automatically productive merely because they occur. Learners may need feedback, prompts, knowledge or enough emotional room to process them. (Narciss & Alemdag, 2025)

The relevant support may not be a cheerful message declaring that mistakes are good.

It may be a much more precise conversation:

Which part was working?

Where did the answer change direction?

What did this attempt reveal?

What could you check next?

When familiar support changes

Sometimes capability appears to disappear when a prompt is removed, the example looks different or the person who usually helps is not present.

That does not necessarily mean no learning occurred. It may show that the earlier performance depended on conditions that have now changed.

The next decision is not automatically to restore every support or remove it all.

It is to notice what the changed condition has revealed.

Choose support for its function, not its format

Labels such as visual tool, micro-content, multimodal resource or anxiety-friendly worksheet describe how something is presented. They do not tell us what it helps the learner do.

A more useful way to choose support is to begin with its function.

What seems to be happening	What support might make possible
The task feels undefined	Make the first action, endpoint or boundaries visible
Too much must be held in mind	Externalise steps, vocabulary or intermediate information
Important foundations are missing	Teach, revisit or connect the prerequisite knowledge
The process is unfamiliar	Model a worked example and make the decisions within it visible
Retrieval causes the learner to stop	Begin with supported retrieval and adjust the cues over time
Errors feel globally revealing	Keep feedback specific to the task, evidence and next decision
Evaluation consumes attention	Separate early practice from public or high-stakes performance
A small variation makes the method unrecognisable	Compare examples and practise identifying what has changed
Help is required before every action	Notice which decisions can now be returned to the learner
The demand exceeds the learner’s present resources	Reduce, defer, renegotiate or change the conditions

The same format can serve several functions.

A concept map might externalise relationships that are difficult to hold in mind. It might help a learner organise what they already know. It might reveal that the relationships have not yet been understood.

But a concept map is not useful simply because a learner is supposedly “visual.” Research does not support matching instruction to fixed learning styles. Combining words and relevant visual representations can help when each contributes to understanding, but adding another mode can also create more information to manage.

Likewise, making content shorter does not automatically make it more learnable. A small card can isolate one useful idea. It can also remove the relationships that give the idea meaning.

The tool should be chosen for the change it is intended to produce.

Support can reveal capability—and conceal it

Suppose a learner completes a multi-step problem with a worked example beside them.

What have we learned?

Perhaps the example made the structure visible and the learner understood how each step followed from the last.

Perhaps they copied each step without knowing why it belonged.

Perhaps they understood the process but used the example to avoid the uncertainty of attempting the first step alone.

Perhaps the example supplied information they would legitimately have access to in the real situation.

The completed task does not distinguish between these possibilities.

This is why it helps to ask:

What did the learner do?

And:

What did the support do?

The answer is not intended to diminish supported performance. Completing something with help is still evidence. It shows what becomes possible under those conditions.

But it is different evidence from completing it independently, explaining why the steps work or recognising how to proceed when the example changes.

The purpose of noticing the difference is not to hurry the learner away from help.

It is to avoid asking the finished product to tell us more than it can.

Feeling calmer and becoming more capable are not the same change

A tool may reduce immediate anxiety because it makes the task clearer, supplies missing knowledge or brings the challenge within reach.

That reduction matters. A learner who can now enter the task has access to an experience that was previously unavailable.

But a tool can also reduce anxiety by removing the moment in which the learner expects difficulty.

The answer remains visible, so it never has to be retrieved.

An adult confirms every step, so uncertainty never has to be carried.

The learner only answers questions they already know they can solve, so an error becomes increasingly unfamiliar and threatening.

The work remains private, so the learner never discovers what happens when another person sees an imperfect attempt.

In each case, the learner may appear calm and successful.

That does not make the support wrong. It means calmness alone cannot tell us what is changing.

Research on clinical anxiety distinguishes between support that enables someone to approach a feared situation and safety behaviour that can interfere with new learning by appearing to prevent the feared outcome. Reviews of exposure-based treatment for children and adolescents suggest that reducing some safety behaviours may improve outcomes, although the evidence is still developing. (Plaisted et al., 2021)

That clinical research should not be turned into a rule that educational supports must be withdrawn or that anxious learners should simply be pushed into distress. A parent or teacher is not conducting exposure therapy, and classroom anxiety may have many causes besides an anxiety disorder.

But it raises a useful educational question:

Is this support helping the learner participate in the difficult part—or ensuring that the difficult part never occurs?

Sometimes preventing it is appropriate. The demand may be unnecessary, unsafe or far beyond what the learner can currently manage.

Sometimes the support provides the first workable way into the learning.

And sometimes the difficult part is closely connected to the capability the learner wants or needs to develop.

We cannot tell which situation we are in merely by observing that anxiety decreased.

Scaffolding should respond to what becomes possible

Scaffolding is often described as temporary support that allows a learner to complete something they cannot yet do independently. Effective scaffolds are closely targeted to the area of need and adjusted as the learner’s capability develops. They may be verbal, visual or written. (Education Endowment Foundation)

“Temporary,” however, does not mean removing support according to a timetable.

It means the support has a relationship with what the learner can currently do.

A complete worked example may become a partially completed one. A direct instruction may become a prompt. Several prompts may become a short checklist. A parent who once sat beside the learner may remain available nearby. A rehearsal may move from private to one trusted person before it moves to a larger audience.

The change need not always be toward less support.

When the task becomes more complex, unfamiliar or consequential, greater support may be needed again. Guidance on worked examples similarly recommends returning to more support when a new procedure or substantial variation increases the demands of the task. (Education Endowment Foundation on worked examples)

The support is responsive rather than linear.

This avoids two unhelpful assumptions:

If the learner still needs a tool, they have not really learned.

And:

If the learner succeeded with the tool, they are ready to do without it.

Neither follows from the evidence.

Independence is not the absence of tools

Many capable people use calendars, checklists, calculators, templates, reference materials, speech-to-text software, colleagues and written procedures.

Their capability does not reside solely in what they can produce without assistance.

It includes knowing:

• which tool fits the situation
• how to use it
• when its output can be trusted
• what to do when it is unavailable
• when the situation exceeds what the tool can support
• when another person’s knowledge is needed

For an anxious learner, a tool should not be treated as suspect merely because it provides continuing support.

The question is whether the tool belongs within the destination.

If a learner is working toward travelling independently and will always have access to a timetable app, using the app may be part of the capability.

If they need to speak from memory in a particular assessment, notes may support preparation but will not fully reproduce the required conditions.

If an employee will use a checklist during the real safety procedure, practising with the checklist may be more authentic than proving they can remember every step without it.

“Doing it myself” does not have to mean doing it alone, unaided and from memory.

It means understanding what successful participation actually requires—and recognising which parts belong to the person, the support and the conditions around them.

The learner may not need a learning tool

It is easy to assume that the learner’s distress should be met with a better resource.

But sometimes the difficulty is not inside the material.

The deadline may be unreasonable.

The learner may be exhausted.

The environment may be noisy, unpredictable or socially unsafe.

The task may require public performance before there has been enough private practice.

The instructions may be contradictory.

The work may be far beyond the foundations that have been taught.

The learner may be carrying consequences that make an ordinary error feel intolerably expensive.

No arrangement of cards, diagrams or encouraging phrases can fully repair those conditions.

The appropriate response may be more time, different teaching, clarification, a quieter setting, a change in how knowledge is demonstrated, professional support—or a decision that the task should not proceed in its present form.

Anxiety should not automatically be interpreted as proof that the demand must be avoided.

It should not automatically be interpreted as a challenge the learner must overcome, either.

Sometimes it is information about the conditions.

How can we tell whether the support is helping?

A tool does not need to eliminate anxiety to be useful.

Nor does a learner need to complete the whole task independently before progress becomes visible.

The first changes may appear around the difficult moment.

The learner begins with one prompt instead of four.

They notice confusion before becoming completely lost.

They can identify which word in the instruction is unclear.

They attempt an answer before asking whether it is right.

They remain with an error long enough to inspect it.

They recover after losing their place.

They recognise that a new example differs from the one they practised.

They ask for more specific help.

They can describe what the checklist is helping them remember.

The anxiety may still be present. The final result may still look uneven. Support may still be needed.

But something within the learning conditions has changed.

A simple observation process can help make that movement visible:

Before the support

• Where did the learner become unable to proceed?
• What seemed to be happening at that point?
• What did the task require them to understand, remember, decide or tolerate?
• Was the expected learning actually accessible?

With the support

• What became possible?
• Which part did the learner now do?
• Which part did the support supply?
• Did the support clarify the learning, reduce an unrelated demand or remove part of the intended capability?

After the attempt

• What remained difficult?
• What did the learner understand that was not visible before?
• Did they recover differently?
• Would the next attempt benefit from the same support, a changed support or different conditions?
• Is the support part of the eventual capability, or are we trying to discover what becomes possible when it changes?

These questions do not produce a universal formula.

They produce better evidence for the next decision.

Supporting the learner without turning anxiety into their identity

There is a difference between saying:

This is an anxious learner.

And:

Anxiety is changing what is available to this learner in this moment.

The first can quietly make anxiety the explanation for everything.

The second leaves room to examine the task, the knowledge, the support, the environment and the conditions under which capability appears.

It also leaves room for variation.

The learner may retrieve the information in conversation but not during a test.

They may write independently about a familiar subject and need extensive structure for an unfamiliar one.

They may explain the process clearly after watching somebody else, then lose it when asked to begin from a blank page.

They may participate with anxiety on one day and be unable to enter the same task on another.

None of those moments needs to become a complete verdict on who they are or what they will eventually be able to do.

The task is not to find the perfect anxiety-friendly format.

It is to keep looking closely enough to ask:

What is difficult here?

What is the learner being asked to develop?

What does this support make possible?

What does it currently do on the learner’s behalf?

What has changed around the difficult moment?

The best support may make the learning feel calmer.

It may make one part clearer.

It may hold information that the learner cannot yet manage while they work on something else.

It may allow anxiety to be present without letting it decide the whole attempt.

And sometimes it may reveal that the most important change is not another tool at all.

Anxiety changes the learning conditions.

When we understand which conditions have changed, we can respond more precisely—without mistaking distress for incapability, supported success for complete independence or immediate comfort for the only evidence that a tool is working.

Research behind this article

Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7(2), 336–353. View abstract

Owens, M., Stevenson, J., Hadwin, J. A., & Norgate, R. (2008). Anxiety and depression in academic performance: An exploration of the mediating factors of worry and working memory. School Psychology International, 29(4), 433–449. View study record

Owens, M., Stevenson, J., Norgate, R., & Hadwin, J. A. (2014). When does anxiety help or hinder cognitive test performance? The role of working memory capacity. British Journal of Psychology, 105(1), 92–101. View abstract

Robson, D. A., Johnstone, S. J., Putwain, D. W., & Howard, S. (2023). Test anxiety in primary school children: A 20-year systematic review and meta-analysis. Journal of School Psychology, 98, 39–60. View abstract

Pashler, H., Bain, P. M., Bottge, B. A., Graesser, A., Koedinger, K., McDaniel, M., & Metcalfe, J. (2007). Organizing instruction and study to improve student learning. Institute of Education Sciences. View practice guide

Education Endowment Foundation. Five a day: Supporting high-quality teaching for pupils with SEND. View guidance

Plaisted, H., Waite, P., & Creswell, C. (2021). Optimising exposure for children and adolescents with anxiety, OCD and PTSD: A systematic review. Clinical Child and Family Psychology Review, 24, 348–369. View abstract

Narciss, S., & Alemdag, E. (2025). Learning from errors and failure in educational contexts: New insights and future directions for research and practice. British Journal of Educational Psychology, 95(1), 197–218. Read article