The English Tutor Handbook Vol No.0011 | Adaptive English Teaching — When to Pause and Fix, Change the Support, or Extend the Learner

The English Tutor Handbook · Volume 0011 · Series ID SETC-ETH-0011

A lesson plan is a prediction.

It predicts what the learner will remember.

It predicts where difficulty will appear.

It predicts how much support will be needed.

Then the learner begins working.

Reality arrives.

Adaptive teaching begins when the tutor treats learner evidence as permission to change the next move.

This volume owns adaptive decision-making inside English tutoring. It does not replace generic adaptive-teaching guidance, the wider eduKate Diagnostic Tutor, Route Designer or Learning Architect, and it does not create separate programmes for every learner preference. Its job is more practical:

When an English learner shows something unexpected, how should the tutor decide whether to pause and fix, change the support, continue, or extend?

Quick Read

• Check before adapting.
• Adapt the next instructional move, not the learner’s identity.
• Use three broad pathways: pause and fix, adapt support, or extend.
• Change one important variable at a time when possible.
• Reduce support when the learner is more capable than predicted.
• Do not respond to every struggle by making the task easier.
• Do not respond to every success by making the task harder.
• Preserve the core learning goal while changing access, sequence, prompt level, representation, practice or load.
• Retest after adapting so the tutor knows whether the change helped.

1. Adaptation Is Not Improvisation Without Evidence

A tutor can change activities constantly and still teach poorly.

Adaptive teaching is not simply variety.

It is a response to evidence.

The Education Endowment Foundation’s 2026 Check. Adapt. materials frame the decision around what learners show: when many misunderstand, pause and fix; when some are unsure, adapt support; when most understand, extend and support. A tutorial has fewer learners, but the decision logic transfers well.

First discover what is happening.

Then change the smallest useful part of instruction.

2. The Three Adaptive Paths

A practical English tutor can begin with three broad paths.

Path A: Pause and fix

The learner’s current model is too weak or inaccurate to make the planned next step useful.

Path B: Adapt support

The learning goal remains appropriate, but the learner needs a different representation, prompt, example, sequence or temporary scaffold.

Path C: Extend

The learner is more secure than predicted. Reduce support, increase discrimination, change context or ask for deeper explanation.

The paths are not labels for students. They are decisions for this learning moment.

3. First Ask: What Evidence Changed?

Do not adapt because the learner says, “This is hard.”

Do not refuse to adapt because the worksheet is “supposed to be at this level”.

Look for evidence.

• repeated error;
• unexpectedly accurate independent work;
• prompt dependence;
• misunderstood task;
• successful oral explanation but failed writing;
• strong untimed work but timed collapse;
• rapid retrieval indicating the current task is too familiar;
• one small cue producing a large improvement.

Adaptation should answer a real discrepancy between the plan and the evidence.

4. Pause and Fix When the Foundation Is Wrong

Some errors are too upstream to ignore.

If a learner has misunderstood what a pronoun refers to, the inference built on top may be meaningless.

If a learner believes a paragraph should contain as many “good words” as possible, continuing into style practice may reinforce the wrong model.

If a learner thinks the task asks for description when it asks for evaluation, polishing the response is wasteful.

Pause.

Repair the foundational relationship.

Then retest before returning to the route.

5. Do Not Pause the Whole Lesson for Every Error

Not every mistake deserves a route change.

A low-cost article error during a paragraph-architecture lesson may be marked for later.

A one-off vocabulary slip may not require vocabulary remediation.

The adaptive question is:

Will continuing under the current plan produce useful learning, or will it rehearse the wrong model?

6. Adapt Support Before Changing the Goal

Sometimes the learner can reach the intended goal if the support changes.

Keep the goal.

Change the access.

• shorter text before longer text;
• oral explanation before written production;
• two evidence candidates before independent location;
• paragraph purpose supplied before learner-generated planning;
• vocabulary cluster before free retrieval;
• untimed execution before time compression.

Support is temporary architecture, not a lowered definition of success.

7. Change Representation

A learner may understand the idea in one representation and fail in another.

Change representation to separate conceptual understanding from language-production load.

• say it before writing it;
• turn a paragraph into a flow of claims and evidence;
• convert a long sentence into two relationship statements;
• map a narrative sequence visually;
• compare spoken and written versions;
• turn listening notes into a timeline.

If the learner succeeds in the alternate representation, the tutor has learned where the bottleneck may sit.

8. Change Prompt Level

Prompt level is one of the cleanest adaptive variables.

If the learner stalls, add the smallest useful cue.

If the learner succeeds too easily, remove a cue.

The least-help-first principle keeps the learner’s thinking visible and prevents over-support.

EEF’s 2026 scaffolding materials emphasise carefully reducing support as learners become more capable, so external scaffolding can become increasingly internal self-regulation. The same principle is highly practical in tutoring.

9. Change Sequence

The planned activity may be appropriate but premature.

For example, a full essay may be the correct eventual task but a poor next task if paragraph architecture is still unstable.

Adapt sequence:

paragraph role → development pair → full paragraph → mixed paragraph set → essay.

The destination remains. The route changes.

10. Change the Example, Not the Concept

Sometimes the example is the problem.

An unfamiliar topic may hide a known reading skill.

A culturally distant scenario may add unnecessary interpretation load.

A sentence may contain vocabulary so difficult that grammar cannot be observed cleanly.

Choose a cleaner example to teach the mechanism, then return to the difficult surface for transfer.

11. Change the Load

A learner may know the operation but fail when too many operations compete.

Reduce load temporarily.

• one paragraph instead of a full essay;
• one evidence relationship instead of five questions;
• shorter listening segment;
• controlled vocabulary range;
• one editing category at a time.

Then reintegrate.

Adaptation should support learning the target, not permanently avoid integration.

12. Change the Time Condition

Timed failure does not always mean the learner lacks the skill.

Untimed success does not always mean the learner is performance-ready.

Use time adaptively.

• remove time to inspect the method;
• restore soft timing once the method is stable;
• use checkpoints rather than one total time;
• train recovery after a time-consuming item;
• reduce perfection demands when completion is the performance target.

13. Extend When the Learner Is More Secure Than Expected

Adaptation also means refusing to waste learner time.

If the learner retrieves the method independently, do not require five more identical examples merely because they were printed.

Extend by changing:

• surface topic;
• question wording;
• genre;
• evidence distribution;
• support level;
• time load;
• requirement to explain the boundary or exception.

Extension should deepen or broaden the capability, not simply add harder vocabulary for prestige.

14. Extension Through Explanation

One powerful extension is to ask the learner to explain why the method works and when it would fail.

This shifts from execution to principle.

A learner who can solve an inference question may be asked to create a near-miss distractor.

A learner who can write a coherent paragraph may be asked to diagnose a weaker paragraph.

A learner who can choose vocabulary accurately may be asked to compare register or connotation.

15. Extension Through Transfer

Another extension is portability.

Can the learner use the same relationship in another mode?

Claim and evidence in reading → claim and evidence in writing.

Changed information in reading → changed information in listening.

Audience awareness in situational writing → audience awareness in oral response.

SETC-ETH-0008 owns the Unseen Transfer Test. Adaptive teaching uses those transfer dimensions as an extension route when current performance is secure.

16. Adaptive Reading Instruction

When reading goes wrong, adapt according to the layer.

• Access problem: reduce lexical or syntactic load temporarily.
• Meaning problem: map people, events, claims or paragraph relationships.
• Reasoning problem: compare interpretations.
• Evidence problem: isolate candidate evidence.
• Answer problem: separate understanding from response construction.
• Performance problem: restore timing only after the method is visible.

The same “wrong answer” should not trigger the same support.

17. Adaptive Vocabulary Instruction

If the learner knows the definition but cannot use the word, do not reteach the definition.

Change the practice to deployment.

If the learner uses the word but misjudges register, compare contexts.

If retrieval fails, remove recognition cues.

If the learner succeeds easily, introduce near synonyms and collocation decisions.

Adaptive vocabulary teaching changes the next operation rather than endlessly adding more words.

18. Adaptive Writing Instruction

Writing produces many visible problems at once.

Adapt by choosing the highest-leverage current layer.

• ideas absent → build content;
• ideas present but scattered → organise;
• structure secure but expression weak → sentence control;
• expression secure but repetitive → development and variation;
• draft strong but editing poor → verification;
• untimed work strong but exams weak → performance conditions.

Do not give the same feedback package to all six conditions.

19. Adaptive Editing Instruction

If the learner cannot detect an error, marking the location changes the task.

That may be appropriate for teaching classification or repair.

Later, remove the location cue to test detection.

Adapt based on which stage fails:

detect → classify → repair → verify.

20. Adaptive Speaking Instruction

Speaking can fail through ideas, organisation, retrieval, sentence control, listener awareness or anxiety under real-time conditions.

Change one support:

• give planning time;
• reduce planning time;
• provide one example;
• remove prepared wording;
• use an unexpected follow-up;
• let the learner speak from bullet points;
• ask for the same idea in simpler language.

The goal is flexible communication, not one rehearsed performance.

21. Adaptive Listening Instruction

If the learner misses everything, replaying the same audio without changing the listening purpose may not help.

Adapt:

• first listen for structure only;
• second listen for one information category;
• provide a note frame temporarily;
• highlight changed information;
• reduce segment length;
• later restore full-length audio and remove the frame.

22. Adaptation in a Three-Student Tutorial

Small groups require adaptation without turning the room into three unrelated private lessons.

Keep the shared learning object when possible.

Vary:

• prompt level;
• response complexity;
• evidence load;
• time allowance;
• independence requirement;
• extension task.

One student may receive candidate evidence.

One may locate it independently.

One may create a distractor and explain why it is weaker.

Same conceptual room. Different support.

23. Adaptation Should Not Become Permanent Differentiation

A temporary scaffold can quietly become a permanent track.

The learner who once needed simplified text may continue receiving simplified text long after the need changed.

The learner who once needed sentence starters may become attached to them.

Every adaptation needs a review condition.

What evidence will allow us to reduce or remove this support?

24. Do Not Adapt to Preference Alone

A learner may prefer short texts, oral work, colourful notes or familiar topics.

Preference can inform engagement.

It should not automatically define instruction.

If the real performance demand requires sustained reading, unfamiliar topics or independent writing, tutoring must eventually prepare the learner for those conditions.

Adaptation is a bridge, not a promise that all future tasks will resemble the learner’s favourite format.

25. Do Not Adapt Difficulty Down Before Diagnosing

When a student struggles, tutors often make the material easier.

Sometimes the problem is not difficulty.

It may be task interpretation.

It may be one unknown word.

It may be missing evidence.

It may be time pressure.

A smaller prompt may preserve the challenge better than a simpler task.

26. Do Not Adapt Difficulty Up Just Because the Learner Is Fast

Fast completion can mean mastery.

It can also mean shallow processing.

Before extending, ask for explanation, boundary or transfer.

A learner who finishes quickly but cannot justify the answer may need deeper verification, not a harder worksheet.

27. Maren: Adapt From Surface to Architecture

Maren’s writing looks strong, so a generic “improve vocabulary” route would be maladaptive.

The tutor notices repeated paragraph redundancy.

Adapt by removing sentence-level polish from the lesson target and shifting to paragraph jobs, deletion and movement.

28. Iona: Adapt Support Without Solving the Evidence Problem

Iona has two relevant evidence candidates and stalls.

Do not simply tell her which one is correct.

Adapt support by giving the exact question relationship and asking her to compare both candidates against it.

The decision remains hers.

29. Leonie: Adapt Conditions, Not Content

Leonie understands the content but performance collapses under time.

More explanation is not the immediate need.

Adapt timing, pacing gates, recovery rules and verification priority.

The knowledge remains. The performance environment changes.

30. Retest the Adaptation

An adaptation is a hypothesis.

After changing the support, ask whether the intended process improved.

If an oral explanation before writing unlocks the paragraph, expression load may be part of the problem.

If candidate evidence does not help, the problem may sit earlier in passage meaning.

If removing time produces no improvement, time pressure may not be the main cause.

The tutor should learn from the response to the adaptation.

31. Adaptive Teaching Needs a Return to the Common Goal

Temporary adaptations should reconnect to the authentic task.

After using bullet points to stabilise an idea, return to full prose.

After using a shorter passage, return to longer reading.

After untimed repair, restore timing.

After giving evidence candidates, require independent location.

Otherwise the adaptation becomes the destination.

32. Research Anchors

The Education Endowment Foundation’s 2026 Check. Adapt. work emphasises checking learner understanding and then using evidence to choose whether to fix, adapt support or extend. Its broader adaptive-teaching materials stress that professional judgement, not a fixed formula, determines the appropriate response.

EEF’s 2026 scaffolding and 2025 metacognition guidance also emphasise explicit support that is reduced as learners assume greater responsibility. The US Institute of Education Sciences’ writing and reading resources similarly support assessment-informed instruction, explicit strategy teaching and opportunities for increasing independence.

33. How this guide fits

• The Tutor Handbook Vol No.0004 | Diagnostic Tutor explains the diagnostic tutor’s role.
• The Tutor Handbook Vol No.0005 | Route Designer explains how to sequence a learning route.
• The Tutor Handbook Vol No.0007 | Learning Architect explains how a complete learning system fits together.
• Evidence Before Intervention covers the English evidence chain before intervention.

This guide focuses on live adaptive English instruction: how tutors change support, representation, sequence, load, timing or extension in response to English-specific learner evidence while preserving the intended learning goal.

34. The Adaptive Decision

1. What did the learner show that the plan did not predict?
2. Is the current model wrong, fragile or already secure?
3. Should I pause and fix?
4. Can the goal remain while support changes?
5. Which single variable would clarify the problem?
6. Can I reduce support instead?
7. Would extension improve discrimination or transfer?
8. What evidence will show whether the adaptation worked?
9. When will the temporary support be removed?
10. How will the learner return to the authentic task?

Adaptive English teaching is not the art of making every lesson different. It is the discipline of changing the next move when the evidence makes the original move less useful.

That is Adaptive English Teaching.

Next: The English Tutor Handbook Vol No.0012 | Tutor Calibration — When the Tutor’s Explanation, Prompt, Material or Assumption Is the Problem.

Explore the English Tutor Handbook. Related teaching guides: Vol. 0012 — Tutor Calibration; Vol. 0014 — Difficulty Calibration.