MRCPCH feedback timing: review now or after the block?

Use immediate feedback when you are learning an unfamiliar or unstable distinction and need correction before the same error repeats. Delay feedback until the end of a short question block when you are rehearsing independent retrieval, pacing or sustained decision-making. Whichever timing you choose, inspect the explanation actively and reattempt the principle later: timing alone does not create durable learning.

The useful question is not, "Is immediate or delayed feedback always better?" It is, "What should this practice session train?" Opening every explanation after every item can fragment a realistic question block. Hiding all feedback until much later can leave an avoidable misconception unresolved. A deliberate feedback rule lets the timing serve the session rather than habit.

What immediate and delayed feedback mean

In question-bank research, immediate feedback usually means seeing correctness, the correct response or an explanation after an individual answer. Delayed feedback may mean after several intervening questions, after the whole test, hours later or even days later. Those are not equivalent delays.

For MRCPCH revision, a practical distinction is:

• item-level feedback: review immediately after submitting one question;
• block-level feedback: complete a bounded set, then review it in the same session; and
• remote feedback: leave the review until a later session or day.

This article mainly compares the first two. A delay of ten questions still preserves the original reasoning reasonably well, whereas a delay of several days risks losing what you noticed, considered and nearly selected. The evidence base also contains comparatively few multi-day delays, so it should not be stretched into a claim that postponing review for days is generally beneficial.

RCPCH currently states that the MRCPCH theory examinations use single best answer questions: one question, five alternatives and one best response. More than one option may have valid elements, so feedback should do more than reveal a letter. It should help the learner identify why one response is superior for the task asked. The PaedVault MRCPCH exam guide summarises the current structure and links to the official pages candidates should recheck near their sitting.

What the evidence says about feedback timing

The strongest current answer is more restrained than many revision slogans.

A 2026 meta-analysis examined 51 computer-assisted learning studies and 160 effect sizes. It found no statistically significant average advantage for immediate or delayed feedback. Effects varied substantially, and educational level, learning domain and response-time constraints were among the moderators. The authors also noted that "delay" ranged from seconds to seven days or from one to 60 intervening items, with limited evidence for delays of a day or longer.

That result matters because it challenges two tempting absolutes: "instant feedback always prevents errors" and "delay always creates a desirable difficulty". Neither is a safe universal rule.

Earlier evidence helps explain the mixed picture. A 1988 meta-analysis of 53 studies found that applied classroom studies often favoured immediate feedback, while experimental studies involving acquisition of test content often favoured delay. A major review of formative feedback later concluded that the findings were inconsistent and recommended matching timing to the desired outcome, task and learner.

Multiple-choice research gives a more specific lesson. In a laboratory study by Butler and Roediger, both immediate and delayed feedback improved later correct recall and reduced intrusion of incorrect alternatives compared with no feedback. This supports providing corrective feedback after multiple-choice practice, but it does not establish one timing as universally superior for MRCPCH preparation.

A systematic review of formative assessment and feedback in higher education also found an unclear immediate-versus-delayed picture in its higher-quality studies. The review emphasised the diversity of tasks, designs and feedback content. A medical-student review found feedback generally beneficial across included randomised trials, but the interventions and outcome timing were heterogeneous and much of the assessment was immediate. It cannot supply a validated timing rule for postgraduate MRCPCH question-bank use.

For MRCPCH revision, feedback presence, accuracy and processing are better-supported priorities than a universal minute-by-minute timing rule. Choose timing to protect the kind of practice you intend to do.

No direct trial identified for this article compared immediate with block-delayed feedback in MRCPCH candidates or measured MRCPCH pass rates.

Choose the timing from the session goal

Start by naming the mode of practice. "Doing questions" is too vague: the same interface can be used to learn, diagnose or rehearse.

Use immediate feedback in learning mode

Immediate feedback is a sensible default when:

• the topic or question form is new;
• your answer depends on an uncertain rule;
• you are repeatedly choosing the same kind of distractor;
• the explanation is needed to understand the next item; or
• the goal is careful acquisition rather than uninterrupted performance.

Commit to an answer and a brief reason before opening the explanation. Otherwise, instant access can turn retrieval practice into assisted recognition. After feedback, state the corrected distinction without copying the paragraph.

The sequence is:

1. answer without notes;
2. record a one-line reason or confidence label;
3. reveal the feedback;
4. identify the decisive correction; and
5. close the explanation and retrieve that correction once.

The confidence-calibration guide explains why a correct but low-confidence response can still deserve feedback, while the self-explanation guide shows how to turn the decisive comparison into your own words.

Use block-level feedback in rehearsal mode

Delay feedback until the end of a short block when the goal is to practise:

• maintaining a decision process across several questions;
• moving on without external confirmation;
• selecting an answer under a time budget;
• recovering attention after a difficult item; or
• measuring a cleaner baseline before targeted repair.

Block mode removes the small correctness signal that can influence the next response. It also makes the session feel more like an assessment than a tutorial. Keep the delay bounded: review the block while your reasoning notes and uncertainty are still interpretable.

Block-level feedback does not mean glancing at a percentage and ending the session. The learning phase begins when the block is marked. Review incorrect answers, correct guesses and unstable final-two decisions, not only the red items.

The MRCPCH answer-changing guide gives a separate rule for reconsidering an answer during a block: change it when a specific reason overturns the first rationale, not merely because uncertainty feels uncomfortable.

Avoid remote feedback as a default

Leaving explanations until tomorrow may be unavoidable after a shift, but it is a poor routine if you no longer remember why each option was plausible. If review must be postponed, preserve a minimal trace before stopping:

• chosen option;
• confidence: high, medium or low;
• nearest alternative; and
• one-line reason.

Then begin the next session with review rather than another large block. Do not interpret the evidence for some delayed conditions as permission to accumulate an ever-growing queue of unreviewed questions.

Use the MODE decision rule

The MODE rule converts the evidence into a practical choice without pretending that it is a validated MRCPCH intervention.

M - Main goal

Is the session teaching a weak distinction or rehearsing independent performance?

• Teach: prefer item-level feedback.
• Rehearse: prefer block-level feedback.

Do not mix modes accidentally. A "mock" with an explanation after every item is a tutorial; a learning session with all correction hidden for hours is an assessment.

O - Original response

Commit before feedback. Record enough of the original decision to prevent hindsight from rewriting it. A confidence label and one sentence are usually sufficient.

If you selected the correct answer for the wrong reason, feedback should correct the reason. If you selected an incorrect answer from a coherent but false rule, preserve that rule long enough to replace it explicitly.

D - Delay boundary

Define the delay before starting:

• after each item;
• after 5 questions;
• after 10 questions; or
• after one planned timed block.

An explicit boundary prevents "delayed feedback" becoming "feedback skipped". For a demanding topic, shorten the block. For later-stage rehearsal, lengthen it gradually while retaining enough time for review.

E - Engage and return

Feedback must change what happens next. For each selected item:

1. compare your original reason with the explanation;
2. isolate the smallest corrected principle or distinction;
3. verify substantive claims against the cited, current source;
4. retrieve the correction without looking; and
5. schedule a changed reattempt after a delay.

The spaced-repetition guide explains how to bring selected material back without converting every question into a permanent card.

A two-lane PaedVault workflow

Use two clearly labelled lanes across the week.

Lane	Purpose	Feedback timing	Evidence of a useful session
Learn	Build or repair a distinction	After each item or a mini-set of 3 - 5	You can state and re-retrieve the corrected rule
Rehearse	Practise independent decisions and pacing	After a planned block	You complete the block without correctness cues, then diagnose the pattern

An illustrative week might include three Learn sessions and one Rehearse session. This is a planning example, not an evidence-based dose. Adjust it to the examination, time available, prior knowledge and what representative results show.

Example Learn session: 30 minutes

• 2 minutes: choose one narrow weakness.
• 15 minutes: answer with item-level feedback.
• 8 minutes: process three high-value corrections.
• 3 minutes: retrieve those corrections with the explanations closed.
• 2 minutes: schedule changed reattempts.

Example Rehearse session: 45 minutes

• 2 minutes: define the question count and time limit.
• 25 minutes: complete the block without feedback.
• 3 minutes: mark and classify by correctness and confidence.
• 12 minutes: review the highest-value decisions.
• 3 minutes: record the next repair target.

If shifts make a continuous block unrealistic, the rota-proof revision plan can help you choose a smaller viable unit rather than abandoning structured practice.

Process explanations instead of collecting them

The content of feedback matters as much as when it appears. A correctness tick answers, "What happened?" A useful explanation should also help answer:

• What task did the lead-in set?
• Which cue or relationship was decisive?
• Why was the nearest alternative inferior here?
• What error produced my choice?
• What changed question would test the same distinction?

Use one short output after review:

"I chose [option/rule] because [original reason]. The explanation changes that reasoning because [decisive correction]. I will recognise the boundary next time by [cue]."

Do not create an unverified clinical rule from memory. PaedVault feedback timing is the subject here; medical propositions remain dependent on current, authoritative sources linked by the question.

Prevent seven feedback-timing failures

Opening the explanation before committing

This removes the diagnostic value of the attempt. Make a choice first, even in immediate mode.

Treating immediate feedback as passive reading

Fast access is not the same as learning. Close the explanation and reconstruct the correction.

Letting feedback interrupt every rehearsal item

If the goal is sustained independent performance, use a block boundary and protect it.

Delaying so long that the reasoning disappears

A score without the original decision process is hard to repair. Preserve confidence and a brief reason, or review sooner.

Reviewing only wrong answers

Correct guesses and low-confidence correct answers can hide unstable knowledge. Use correctness together with confidence.

Copying the whole explanation

Transcription creates an archive, not a retrieval cue. Extract the decisive distinction and test it again.

Mistaking a block score for a readiness verdict

One percentage is sensitive to topic mix, difficulty and chance. Look for patterns across representative blocks and use feedback to decide the next action.

Measure whether your timing rule works

Run a two-week personal audit without claiming it is an experiment.

Track only a few signals:

• proportion of planned blocks actually reviewed;
• repeated errors from an already reviewed distinction;
• correct answers originally marked low confidence;
• ability to retrieve selected corrections after several days; and
• whether item-level feedback is fragmenting planned rehearsal.

If immediate mode produces careful corrections but you struggle to finish independent blocks, add a rehearsal lane. If block mode produces many vague notes and repeated misconceptions, shorten the block or return to item-level learning for that topic.

Do not compare unlike question sets and attribute every score movement to feedback timing. The audit is a way to improve workflow, not proof of causation.

Conclusion

There is no evidence-based universal winner between immediate and delayed feedback for MRCPCH revision. The current research suggests that timing effects depend on the learner, task, domain and definition of delay.

Use immediate feedback to teach and repair. Use short block-delayed feedback to rehearse independence. Avoid unbounded postponement. In either mode, commit first, examine the explanation actively and retrieve the corrected distinction again later.

Frequently asked questions

Should I review MRCPCH feedback after every question?

Use after-each-question feedback when learning an unfamiliar or unstable distinction. For timed practice or sustained decision rehearsal, complete a short planned block first. Always commit to an answer before revealing feedback, then process selected explanations actively rather than reading them passively.

Is delayed feedback better for memory?

Not universally. A 2026 meta-analysis of computer-assisted learning found no significant average difference between immediate and delayed feedback, with substantial variation across contexts. Some individual studies favour delay, while classroom studies have often favoured immediacy. The practical choice should match the learner, task and session goal.

How long should I delay question-bank feedback?

No validated MRCPCH interval was identified. A useful block delay might be 5 - 10 questions or one planned timed set, followed by review in the same session. Define the boundary in advance and avoid postponing so long that you forget the reasoning behind your choices.

Should I review correct MRCPCH answers?

Review correct answers that were guesses, low confidence or based on a reason contradicted by the explanation. Correctness alone can conceal an unstable decision. Secure, high-confidence answers usually need less immediate attention, leaving time for higher-value corrections and later mixed retrieval.

Sources

• RCPCH theory exam structure and syllabi
• A meta-analysis of the impact of feedback timing on learning outcomes in computer-assisted learning
• Timing of feedback and verbal learning: a meta-analysis
• Focus on formative feedback
• Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing
• Formative assessment and feedback for learning in higher education: a systematic review
• The use of feedback in improving the knowledge, attitudes and skills of medical students: a systematic review and meta-analysis of randomised controlled trials