The Tutor Handbook Vol No.0072 | The Feedback Receipt

The Tutor Handbook · Volume 0072 · Series ID THB-0072

The Tutor Handbook: Complete series index. For the wider mechanism of feedback in teaching, see How Feedback Works in Teaching. This volume owns a narrower tutoring decision: how the tutor verifies whether feedback actually changed the learner’s next move.

A learner nods and says, “Yes, I understand.” The tutor has just explained why a comprehension answer was too broad. The learner can repeat the tutor’s explanation almost word for word. The corrected answer now looks excellent. The lesson feels successful.

Then a fresh question arrives ten minutes later. The learner writes another broad answer.

The problem was not that the feedback was unheard. It was heard. It may even have been understood. The problem is that the feedback did not yet become a better decision at the moment the learner needed to act again.

The Feedback Receipt is the evidence that feedback changed the learner’s next relevant action under conditions that still require the learner to think.

This is a different standard from agreement, appreciation, correction-copying or short-term recognition. A learner can sincerely agree with feedback and still be unable to use it. A learner can produce a perfect correction while the tutor is guiding every step. A learner can explain the rule immediately after hearing it and then fail to notice when the same rule matters in a changed context.

Feedback therefore has two lives. The first is communicative: did the learner receive the information? The second is operational: did the information reorganise what the learner did next?

The direct answer

A tutor should judge feedback by what happens on a later relevant attempt, not by the learner’s reaction at the moment of explanation. The cleanest sequence is: preserve the first attempt, give feedback that identifies the useful change, ask the learner to reconstruct the change in their own words, then require a new attempt in which the learner must notice when and how to use the feedback. After that, return again later or under changed conditions. If the improvement survives, the feedback has a stronger receipt. If it disappears, the feedback may have produced recognition without durable change.

The Education Endowment Foundation’s feedback synthesis describes feedback as information about performance relative to a learning goal and emphasises that effective implementation includes opportunities for pupils to act on it. That distinction matters here: information is not yet uptake. The tutor needs an opportunity to observe action after the information arrives. The EEF also notes that feedback effects vary by context and that implementation matters, so a tutor should not treat “I gave feedback” as a sufficient explanation for later improvement.

1. What this volume owns

This volume owns the receipt after feedback. It does not own the general design of feedback; that broader job belongs to How Feedback Works in Teaching. It does not own generic error correction; How Learning From Mistakes Works explains the wider learning mechanism. It does not own the learner-facing handoff from a returned mark or comment into an executable study action; the existing Feedback Handoff owns that interface job.

The Tutor Handbook question is practical and narrow: after feedback has been given, what evidence justifies saying that it worked for this learner on this learning job?

2. Agreement is not a receipt

Learners often agree with feedback for good reasons. The explanation is clear. The error becomes obvious once someone points it out. The learner may feel relief because the confusing part finally has a name. Agreement can therefore be genuine and valuable.

But agreement is evidence about the conversation, not yet about later performance. “Yes, I see why” can mean several different things: I recognise the better answer when it is shown; I understand the explanation while the example is still present; I can repeat the rule; I know what the tutor wants me to say; or I can independently detect and repair the same kind of problem in a fresh task. Those states overlap, but they are not identical.

The tutor should therefore avoid converting social smoothness into a learning claim. A polite, attentive learner can still need another attempt. A frustrated learner can still have understood the feedback. The next task is a better receipt than the emotional tone of the explanation.

3. A correction is not a receipt

A corrected answer can be misleading because the learner may be operating in a heavily supported condition. The original task is familiar. The tutor has just named the error. The correct route is active in working memory. The learner may be editing rather than deciding.

That does not make correction useless. Correction is often the first place where the new route is assembled. It lets the tutor check whether the learner can apply the explanation while support is still nearby. The mistake is to stop there and call the repair complete.

A stronger receipt requires some distance from the correction: a fresh example, a changed surface, a delay, less prompting, or a task where the learner has to notice that the feedback is relevant without being told.

4. The feedback chain has four stages

For tutoring, it is useful to separate four stages that are often collapsed into the phrase “the student understood the feedback.”

Reception: the learner notices and can attend to the feedback.
Interpretation: the learner can explain what the feedback means and why it matters.
Execution: the learner can use the feedback on a relevant new attempt.
Transfer and return: the learner can still use the underlying decision later or when the surface changes.

A tutor can observe failure at any stage. If the learner never understood the comment, a new attempt will not help until the meaning is repaired. If interpretation is sound but execution fails, the learner may need guided practice. If execution succeeds only on near-identical examples, transfer is not yet established. If success disappears after a delay, the feedback may have created an immediate correction rather than a stable change.

5. Preserve the original attempt

Before feedback changes the work, preserve enough of the original attempt to remember what happened. In a Mathematics problem, keep the original representation and first method choice. In an English answer, keep the first claim and evidence selection. In Science, keep the original causal explanation rather than rewriting it into polished model language before the tutor has identified the learner’s actual reasoning.

This matters because memory rewrites mistakes quickly. Once the correct answer is visible, both tutor and learner may remember the first attempt as more obviously wrong than it felt at the time. The preserved attempt protects the diagnosis from hindsight.

Volume 0050, The Decision Record, handles this principle at route level. Here the same discipline is used at feedback level: preserve the before-state so the after-state remains interpretable.

6. Name the learner decision, not just the error

“Too vague.” “Careless.” “Explain more.” “Check your answer.” These comments can be true while leaving the learner without an executable next move. A useful feedback receipt begins with feedback that points to a decision the learner can reproduce.

For example, instead of “your evidence is weak,” say: “The line you chose is related to the paragraph, but it does not directly support the inference. On the next question, compare the two most relevant lines and choose the one that would still support your answer if the other line disappeared.”

Instead of “be careful with signs,” say: “The sign changed when you expanded the bracket. On the next question, stop immediately after expansion and check every term against the sign outside the bracket before continuing.”

The feedback now has an action that can generate a receipt.

7. Ask the learner to reconstruct the feedback

After feedback, ask a short reconstruction question: “What will you do differently on the next one?” This is not a memory quiz about the tutor’s wording. The learner should state the operational change in their own language.

If the learner says, “I need to be more careful,” the feedback may still be too vague. Ask: “What will careful look like at the exact point where the error happened?” The aim is to move from an attitude word to a decision rule.

Do not overinterpret a good verbal reconstruction. A learner can explain a rule better than they can use it. The reconstruction is a bridge between interpretation and execution, not the final receipt.

8. The first receipt is a fresh next attempt

The most useful immediate check is usually a fresh task that preserves the relevant reasoning job while changing enough surface detail to require a real decision. If the original algebra problem involved expanding a negative bracket, the next problem should not simply repeat the same numbers. If the original comprehension answer used broad evidence, the next passage should require another evidence choice rather than asking the learner to rewrite the same answer.

This fresh attempt answers a concrete question: can the learner use the feedback when the correct answer is no longer the obvious edit target?

Success here is encouraging. Failure here is also useful because it tells the tutor the feedback has not yet become an executable routine. The tutor can now distinguish “heard” from “usable.”

9. Do not announce the feedback category every time

If the tutor says, “Here is another sign-error question,” the learner no longer has to detect the relevant feature. If the tutor says, “Use the evidence rule we just discussed,” the task has already supplied part of the decision.

Early practice can legitimately contain cues, especially when the learner is still assembling the new route. But a stronger receipt comes when the task itself—not the tutor’s label—must trigger the relevant decision.

This is where the feedback problem touches transfer. The learner has not fully used feedback if they can only act when someone announces which feedback item is currently being tested.

10. Same-surface success can be a false positive

Imagine a learner who has been corrected on percentage base. Immediately afterwards, three near-identical percentage questions are answered correctly. The result may reflect real learning. It may also reflect short-term pattern matching while the corrected structure is highly available.

A changed-surface check is more informative. Ask for the same underlying relationship in a word problem, a diagram or an inverse form. If the learner still identifies the correct base, confidence rises that the feedback changed the underlying decision rather than only the local answer form.

Volume 0019, The Changed Question, owns the general changed-surface test. The Feedback Receipt uses that tool for one purpose: verifying uptake.

11. Delayed return is the second receipt

Immediate success tells the tutor what the learner can do while the feedback episode is still recent. A delayed return asks whether the new decision remains available after the conversation has faded from working memory.

The delay does not need one universal schedule. It should be long enough to reduce immediate carryover and practical enough to fit the learning route. For some work, the next lesson is suitable. For a frequently used routine, later in the same lesson plus a later-day or later-week return may be more informative.

The important distinction is not “one day versus three days.” It is immediate correction versus later independent availability.

12. The tutor should watch the first relevant moment

On a later task, do not look only at the final answer. Watch the moment where the learner should use the feedback.

For Beatrice, that may be the moment she chooses which line counts as evidence. For Alicia, it may be the sign check immediately after expansion. For Ciara, it may be the transition from observation to causal explanation. For Emily, it may be the decision to protect retrieval before filling the rest of the study plan.

A final correct answer can hide a wrong route repaired later by luck or secondary checking. Observing the decision point tells the tutor whether the feedback entered the learner’s process where it was meant to.

13. Feedback can improve output without improving independence

The tutor gives increasingly precise prompts. The learner’s work becomes excellent. The tutor concludes that feedback is working.

Possibly. But if the quality depends on the tutor continuing to identify the error, supply the criterion and remind the learner when to use it, the improved output may partly be a product of the feedback system rather than a fully internalised learner capability.

This does not make support illegitimate. Support is often the mechanism through which learning begins. The receipt becomes stronger when some of the detection, selection and correction work moves into the learner’s control.

The relevant question is: who is now performing the target decision?

14. Feedback saturation can reduce uptake

A page covered in comments may contain accurate observations and still be a poor learning instrument. The learner may not know which change matters first. Several comments may compete for the same next attempt. The tutor may correct vocabulary, structure, grammar, evidence and presentation at once, then be unable to tell why the next version improved.

Feedback has an opportunity cost. Attention spent acting on one comment cannot simultaneously be spent acting on every other comment. The EEF implementation guidance explicitly advises considering opportunity cost and ensuring feedback can be acted upon.

A tutor should therefore ask which feedback item has the greatest current learning leverage. Other observations can remain visible without becoming simultaneous intervention targets.

15. Feedback priority should match the first useful weak link

Suppose a learner’s Science answer is factually correct but poorly structured, and the causal mechanism is also incomplete. Which feedback should come first?

If the target is scientific explanation, repairing the missing mechanism may have greater leverage than polishing sentence order. If the mechanism is sound but hidden by an unreadable sentence, language structure may temporarily be the earlier weak link. The correct priority depends on what currently prevents the learner from demonstrating the intended capability.

This connects the Feedback Receipt to Finding the First Weak Link in Learning. Feedback should not merely describe everything wrong. It should change the earliest useful decision the learner can act on.

16. A feedback receipt can fail because the task was badly chosen

The learner receives excellent feedback on inference. The tutor then tests uptake with a question that requires almost no inference because the answer is stated directly. The learner succeeds.

That success does not test the feedback.

Likewise, a learner receives feedback on method selection and is retested with a blocked worksheet where the method is announced by the heading. The task removes the very decision the tutor claims to be testing.

Volume 0068, The Opportunity Check, is essential here. A failed or successful feedback receipt is only meaningful if the learner had a real opportunity to perform the target skill.

17. Mathematics case: Alicia and the sign that keeps returning

Illustrative composite case. Alicia expands −3(2x − 5) as −6x − 15. The tutor points to the second term and asks what multiplication should happen between −3 and −5. Alicia immediately sees the error and corrects the line to −6x + 15.

A weak feedback process ends there: “Good, you understand.” A stronger process asks Alicia to state the decision: “When a negative multiplier sits outside a bracket, I will multiply every term including its sign before I continue.”

The tutor then gives a fresh expression where the negative sign appears in a different position. Alicia expands it accurately without a sign reminder. Later in the lesson, the tutor includes another bracket inside a mixed set without announcing that sign checking is being tested. Alicia stops after expansion, checks each term, and continues.

The receipt is not the corrected first expression. The receipt is the changed routine appearing at the right moment on a fresh task.

18. English case: Beatrice and evidence that is related but not direct

Illustrative composite case. Beatrice writes a comprehension answer claiming that a character is reluctant to enter a room. She cites a sentence showing that the character is standing by the door, but ignores a later line showing the character repeatedly delaying and asking whether anyone else has arrived.

The tutor’s feedback is not “use better evidence.” It is: “When two lines are relevant, choose the one that most directly changes the probability of your interpretation. Ask which line would make your inference hardest to defend if it disappeared.”

Beatrice can repeat the rule. That is interpretation. The tutor then gives a fresh passage with two plausible evidence lines. Beatrice chooses the stronger one and explains why the weaker line is only contextual. A week later, she does the same on school work without the tutor’s phrase visible.

That later independent evidence is a stronger receipt than the polished correction produced during the lesson.

19. Science case: Ciara and feedback that becomes model-answer dependence

Illustrative composite case. Ciara gives an incomplete explanation of why water droplets form on the outside of a cold cup. The tutor shows a strong causal explanation and Ciara copies its language accurately. Her next answer is excellent when the model remains visible.

The tutor now needs to separate language uptake from mechanism uptake. Ciara is asked to close the model and explain a changed case: droplets forming on a cold metal pipe. She must identify the source of the water, the cooling of surrounding water vapour and the change of state.

If Ciara reconstructs the mechanism in her own words, the feedback has travelled further. If she can only reproduce the exact sentence from the model, the output improved but the causal route may still be externally carried.

20. Studying case: Emily and the feedback that disappears on busy weeks

Illustrative composite case. Emily’s tutor notices that she repeatedly allows urgent homework to displace retrieval practice. The tutor advises protecting a ten-minute retrieval floor before flexible homework expansion. Emily agrees and builds a good plan during the lesson.

The feedback receipt is not the plan created beside the tutor. The receipt appears on a later busy evening when Emily independently protects the retrieval floor despite several competing tasks, or makes a reasoned exception because a genuine deadline changes the opportunity cost.

Feedback about studying is especially vulnerable to false receipts because the tutor often sees the plan rather than the later execution. The return evidence must come from what actually happened outside the lesson.

21. Feedback should sometimes be tested without repeating the wording

If every retest uses the tutor’s exact phrase, the learner may become dependent on the phrase rather than the underlying criterion. “Choose the line that most directly changes the probability of your inference” is useful teaching language. Eventually the learner should be able to make the evidence decision without hearing that sentence first.

This is similar to scaffold fading. The language can remain while the idea is new, then shrink as the learner begins to recognise the situation independently. The tutor should fade the prompt, not the criterion.

A strong feedback receipt is therefore often accompanied by less tutor language over time.

22. The learner can become the detector

The mature form of feedback uptake is not merely using a correction after someone points to the error. It is noticing the error class independently.

Ask the learner before tutor review: “What part of this answer are you least confident about?” “Where is the decision most likely to fail?” “Which criterion would you check first?” These questions should not become a compulsory script for every task, but they can reveal whether the learner is beginning to internalise the detection function.

Feedback has travelled furthest when the learner can generate useful feedback for their own next attempt without waiting for the tutor to identify the weak point.

23. Self-assessment can also create false receipts

A learner may confidently say, “I checked it,” while applying the wrong criterion. Self-assessment is not automatically accurate merely because it is independent. The tutor should compare the learner’s self-judgement with performance evidence and help calibrate the criteria being used.

The purpose is not to make the tutor the permanent judge. It is to improve the learner’s capacity to judge their own work against the actual learning goal.

This is why feedback uptake and metacognition intersect. The learner must increasingly know not only what to change, but how to detect whether the change is needed.

24. Feedback timing should serve the learning job

There is no universal rule that feedback must always be immediate or always delayed. Immediate feedback can prevent repeated practice of a harmful error when the learner is still acquiring a new procedure. Delayed discussion can be useful when the tutor needs to observe a complete independent attempt before interrupting it.

The receipt question remains the same regardless of timing: after the feedback arrives, does the learner later make a better decision?

The EEF feedback synthesis explicitly notes that feedback can be effective during, immediately after and some time after learning. Tutors should therefore choose timing based on purpose rather than ritual.

25. Emotional reception matters without becoming the only outcome

Feedback can be accurate and still fail if the learner cannot engage with it. Humiliation, overload or a sense that the comment attacks identity rather than work can reduce the chance that the learner will use the information productively.

This does not mean feedback should be emptied of challenge. It means the tutor should keep the object of feedback clear: the task, strategy, decision or process that can change. Comments about stable personal qualities are less actionable.

A learner can also dislike useful feedback. Emotional discomfort does not prove the feedback is wrong. The tutor should distinguish threat, confusion, overload and productive difficulty rather than treating every negative reaction as a stop signal.

26. Parent feedback can interfere with the tutor’s receipt

A tutor asks the learner to identify evidence independently. At home, a parent points out the correct line before the learner begins. The homework answer improves, but the intended feedback receipt disappears because the key decision was supplied externally.

This does not mean parents should provide no help. The tutor should explain which help preserves the target operation. A parent might ask, “Which line supports your answer most directly?” without selecting the line. If the target is planning, the parent can remind the learner to start the planning process without constructing the plan.

Home support should make the feedback executable without quietly performing the learner’s target decision.

27. Digital and AI feedback needs the same receipt

An AI system can produce detailed comments in seconds. A digital platform can mark every item. The quantity and speed of feedback can be impressive while the learner remains passive.

The tutor should ask the same questions: what exactly did the learner need to notice? Did the learner interpret the feedback? Did the learner make a fresh attempt before seeing another generated answer? Can the learner detect the same problem without the tool identifying it? Does the improvement survive when tool assistance is appropriately reduced?

Digital feedback can be useful. It should not be judged by how much text it produces. The receipt is still changed learner action.

28. Feedback in a three-student tutorial

Three-student tuition creates a useful opportunity: one learner’s feedback episode can become an observation task for the others without turning peers into markers.

Suppose Alicia receives feedback about method selection. Beatrice and Ciara can be asked to identify what changed between the first and second attempt, then return to their own independent work. This can sharpen attention to criteria while keeping the target learner’s route distinct.

But peer exposure can also contaminate the receipt. If the other students announce the correction before Alicia’s fresh attempt, the task no longer measures Alicia’s independent uptake. Group orchestration should therefore separate shared noticing from individual proof.

29. Feedback should not create permanent tutor dependence

A tutoring programme can become extremely good at finding errors for the learner. Every week the tutor detects, explains and repairs. The learner improves while supervised but remains weak at noticing when something has gone wrong.

The final job of feedback is therefore partly to make the feedback provider less necessary. Detection, comparison, checking and repair should increasingly migrate toward the learner where developmentally appropriate.

This links to How Independent Learning Works and to Volume 0040, The Independence Test.

30. A compact feedback-receipt protocol

Preserve: keep the original attempt or enough evidence of the original decision.
Name: identify the smallest useful decision that should change.
Explain: give feedback that makes the criterion and mechanism clear.
Reconstruct: ask the learner what they will do differently, in their own words.
Reattempt: require a fresh task with the same underlying job.
Reduce cues: avoid announcing the feedback category when independence is being tested.
Change surface: use a different representation, wording or context when appropriate.
Return later: check whether the decision remains available after delay.
Move detection inward: ask the learner increasingly to identify the relevant error or criterion themselves.
Update: if the error persists, decide whether the problem is feedback clarity, insufficient practice, wrong diagnosis, poor opportunity to demonstrate, or a different weak link.

31. When the receipt fails

A failed receipt does not automatically mean the learner ignored the feedback. Several explanations remain possible.

The feedback was not understood.
The feedback described the error but did not specify a usable action.
The learner understood the action but has not practised it enough.
The new task removed the opportunity to use the action.
The task added another difficulty that overwhelmed the new routine.
The tutor tested a different skill from the one that received feedback.
The original diagnosis was wrong.
The learner needs a scaffold before independent execution is realistic.
The feedback was too broad or one of several competing priorities.
The learner can execute the change but cannot yet detect when it applies.

The right response is not automatically “give the same feedback louder.” Reopen the relevant stage of the chain.

32. When the receipt succeeds once

One successful fresh attempt is useful evidence but not necessarily stable learning. The result should influence confidence proportionately. A high-consequence route change should demand stronger receipts than a small local adjustment.

For a narrow procedural correction, one fresh attempt plus a later check may be enough to move from active repair into light monitoring. For a complex reasoning habit, the tutor may need several tasks across different contexts before treating the new decision as stable.

The standard is not perfection. It is enough repeated evidence to justify the next educational decision.

33. Feedback can change the tutor too

If a learner repeatedly fails to use feedback, the tutor should not assume the learner is the only part of the system that needs adjustment. The feedback may be poorly timed, too abstract, too numerous, badly matched to the task, or dependent on language the learner does not yet command.

AERO’s Monitor Progress guidance emphasises using student responses to adjust instruction, guidance and feedback. The learner’s next attempt is therefore evidence about the tutor’s feedback design as well as evidence about the learner.

This is an important professional habit: feedback is not something the tutor does to the learner. It is part of a loop in which the learner’s response tells the tutor whether the intervention was usable.

34. Evidence foundation and limits

The EEF Teaching and Learning Toolkit summarises a large feedback evidence base and reports positive average effects, while also stressing implementation, actionable information and opportunities for learners to act. AERO’s Monitor Progress guidance similarly links checking for understanding with responsive teaching, guidance and feedback. These are broad educational sources rather than trials of the exact “Feedback Receipt” protocol described here.

The protocol in this volume is therefore a tutoring application built from established principles: feedback should relate to a learning goal; learners need opportunities to use it; performance conditions matter; repeated and changed-condition evidence strengthens inference; and support should move toward greater learner independence when appropriate.

It would be an overclaim to say that every feedback episode needs the same number of reattempts, the same delay or the same wording. Subject, age, prior knowledge, stakes, accessibility and the learning target all matter.

35. Tutor audit

What exact learner decision am I trying to change?
Did I preserve the original attempt?
Does my feedback identify an actionable criterion rather than a vague judgement?
Can the learner explain the change in their own words?
Did I give a genuine opportunity to act on the feedback?
Was the new task fresh enough to require a decision?
Did I accidentally announce the answer category?
What support remained during the reattempt?
Did the learner detect the relevant moment independently?
Did the improvement survive a changed surface?
Did it survive a later return?
Can the learner increasingly generate useful self-feedback?
If the receipt failed, which stage of the chain should I reopen?

36. Parent audit

After feedback, does my child get a chance to try again?
Am I helping them use the criterion, or am I supplying the decision?
Do corrected answers look better only while the adult explanation is nearby?
Can my child explain what they changed and why?
Does the change appear on fresh school work?
Are there so many comments that no single next action is clear?
Am I treating a polished correction as proof that the problem is solved?

37. Learner audit

What did the feedback tell me to notice?
What will I do differently on the next attempt?
Can I recognise when the same issue appears without someone naming it?
Can I use the idea when the question looks different?
Can I still use it later?
What part still needs a prompt?
Can I identify one useful piece of feedback for myself before the tutor speaks?

38. What the Feedback Receipt is not

It is not the learner saying “I understand.”
It is not a perfect corrected answer produced beside the tutor.
It is not copying model language.
It is not praise, gratitude or compliance.
It is not one high score on an easier task.
It is not a demand that all support disappear.
It is not proof that feedback alone caused every later improvement.
It is evidence that the feedback changed a relevant learner action under conditions that still required the learner to do the thinking.

39. The ethical standard

Feedback creates an unusual asymmetry. The adult can see the better answer before the learner can. That makes it easy to produce impressive work by carrying too much of the correction. It is also easy to confuse compliance with learning because the learner has every social reason to say the explanation makes sense.

A responsible tutor therefore waits for a receipt that belongs increasingly to the learner.

Feedback has done its educational job when the tutor’s information becomes the learner’s better decision, and that better decision can reappear when the tutor is no longer standing beside it.

Evidence and connected reading

Final compression

Preserve the first attempt. Name the decision that should change. Give feedback the learner can act on. Ask the learner to reconstruct the action. Require a fresh next attempt. Reduce the cue. Change the surface when useful. Return later. Watch whether the learner notices the relevant moment. Let detection and repair move gradually into the learner’s control.

Do not call agreement a receipt. Do not call a copied correction a receipt. Do not call a better output a receipt when the tutor still performs the key decision.

The Feedback Receipt is the next independent decision that becomes better because the feedback existed.

That is Tutor Handbook Volume 0072.

Small Group Tutorials

The Tutor Handbook Vol No.0072 | The Feedback Receipt — How a Tutor Checks Whether Feedback Changed the Learner’s Next Attempt Rather Than Only Producing Agreement