Worked examples or practice questions for MRCPCH?

Use both, but change the balance as your knowledge changes. When an MRCPCH question type or reasoning process is unfamiliar, study a high-quality worked example actively: predict each step, explain why it is justified and identify the decisive information. Then remove support gradually and solve a closely matched question without the example. Once you can explain the process and transfer it to new questions, independent retrieval and mixed practice should dominate. Do not keep rereading complete solutions after they have become predictable.

Worked examples and practice questions do different jobs. A worked example makes a successful process visible. An independent question tests whether you can select and execute that process without help. Treating either one as the whole revision method creates a gap: examples can become passive, while unsupported questions can produce repeated guessing.

This guide applies learning research to Foundation of Practice (FOP), Theory and Science (TAS) and Applied Knowledge in Practice (AKP) preparation. It does not provide clinical instruction, and the sequence below is an evidence-informed study framework rather than an RCPCH requirement.

The direct answer: show, fade, solve, then return

Choose the starting point from your current performance on the specific task:

• Unfamiliar process: begin with a fully worked example.
• Partly understood process: complete a partially worked example or predict the missing steps.
• Reliable process: answer independently, then use the explanation as feedback.
• Apparently secure process: test it later in a mixed set without cues.

The important transition is from supported understanding to unsupported performance.

A worked explanation is useful when it changes what you can do on the next unseen question. If it only feels clear while it is open, it has not yet become independent exam knowledge.

RCPCH provides current syllabi to define examinable scope and online sample papers to help candidates understand the computer-based experience. Neither source prescribes a worked-example revision ratio. The practical sequence in this article is therefore an application of broader learning evidence, not an official College formula.

What counts as a worked example?

A worked example is more than an answer key. It displays a problem, a reasoned route through it and a justified conclusion. For MRCPCH preparation, a useful worked example can include:

1. the task being asked;
2. the information that changes the decision;
3. the tempting but less suitable alternatives;
4. the rule, relationship or distinction used;
5. the final choice; and
6. a brief check showing why the choice fits the lead-in.

A paragraph that merely repeats the correct option is feedback, but it is not a strong worked example. Conversely, a very long explanation can conceal the decision beneath facts that were not needed.

The safest study examples use independently authored material mapped to the public syllabus. Official RCPCH sample papers are particularly useful for becoming familiar with the current computer presentation, but they are a limited orientation resource rather than an unlimited question bank.

What the evidence supports - and where it stops

Worked examples can help initial skill acquisition

A 2023 meta-analysis examined 55 studies and 181 effect sizes in mathematics learning. Worked-example conditions produced a medium average benefit over comparison conditions, but the studies were heterogeneous and the authors identified evidence of publication bias. Most studies concerned algebra or geometry rather than postgraduate medical examinations.

That evidence supports a cautious principle: when a structured process is new, seeing a solution can be more efficient than repeatedly attempting it without a model. It does not prove a particular number of examples for MRCPCH, nor does it show that reading explanations alone prepares a candidate for best-of-five questions.

The right amount of help depends on prior knowledge

A 2025 meta-analysis of the expertise-reversal effect included 60 experimental studies and 5,924 participants. On average, learners with lower prior knowledge benefited from higher-assistance instruction, while learners with higher prior knowledge benefited from lower-assistance instruction. Effects varied by domain and other study features.

This matters because "novice" is topic-specific. A paediatric trainee may be secure in one syllabus area and unfamiliar with another. Use recent task performance - not professional seniority, confidence or total revision hours - to decide how much support to provide.

Active processing matters

In a randomised study of 54 third-year medical students, combining self-explanation with examples of residents' reasoning and prompts improved some one-week transfer outcomes compared with the control condition. The study was small and concerned undergraduate diagnostic reasoning, not MRCPCH performance, but it illustrates an important distinction: exposure to an example and active processing of an example are not the same task.

Ask questions while studying the solution:

• What was the first decision?
• Which clue or relationship made that decision possible?
• Why was the nearest alternative rejected?
• Which step would I have missed?
• Can I reconstruct the route with the explanation hidden?

Independent testing remains essential

A study of 47 first-year medical students compared repeated testing and repeated study, with and without self-generated explanations. Six months later, repeated testing produced better retention and application than repeated study; self-explanation also helped, but less. The sample and subject matter limit direct generalisation, yet the finding reinforces the need to move from studying solutions to retrieving without them.

Examples establish or repair a process. Questions show whether the process is available when the support disappears.

Use the SHOW - FADE - SOLVE - RETURN sequence

This is a PaedVault editorial framework, not an RCPCH protocol.

1. SHOW: study one model actively

Choose a worked example that matches the exact difficulty:

• interpreting the lead-in;
• identifying the decisive information;
• comparing two plausible options;
• organising a multistep calculation or mechanism; or
• selecting a strategy under time pressure.

Before revealing each part, make a prediction. After revealing it, explain the link in your own words. Cover the solution and reconstruct the route once.

Avoid copying the explanation verbatim. Transcription can preserve the page without revealing whether you understood the decision.

2. FADE: remove part of the support

Turn the next item into a completion problem. Keep the stem and perhaps the first step, but hide later reasoning. Alternatively, retain the conclusion and reconstruct the decisive route.

Useful fading prompts include:

• "The lead-in asks for ___."
• "The decisive information is ___ because ___."
• "Option B is less suitable because ___."
• "The next step in the reasoning is ___."

Fade support quickly when the missing steps are produced accurately. Restore one level of help when the candidate cannot name the task, the decisive distinction or the reason for the answer.

3. SOLVE: attempt a matched question independently

Use a new question that tests the same process without copying the surface details. Work closed-book, commit to an answer and record confidence before opening feedback.

Then compare process as well as outcome:

Result	Interpretation	Next action
Correct, sound reasoning	Support may be removable	Try a less similar question
Correct, guessed or flawed reasoning	Knowledge is not yet reliable	Return to the decisive step
Incorrect, process understood	A narrower knowledge gap may remain	Repair that gap, then retry
Incorrect, process unclear	Support faded too quickly	Study or complete another example

The best-of-five question strategy provides a consistent independent attempt process. The incorrect-answer review guide helps classify what failed after the attempt.

4. RETURN: test transfer after a delay

A same-session success can reflect short-term familiarity with the example. Schedule a later question that requires the same underlying process but appears in a different topic or format.

On return, do not reopen the original example first. Attempt the question, explain the decision and only then compare with feedback. If the process is unavailable, use the smallest amount of support needed to rebuild it.

The spaced-repetition guide explains how to schedule delayed retrieval without turning every explanation into a flashcard.

Decide where to enter the sequence

Use a brief diagnostic attempt:

1. take one representative question without notes;
2. state what the lead-in requires;
3. name the decisive principle or distinction;
4. choose and justify an answer; and
5. rate whether the route was known, reconstructed or guessed.

Enter at SHOW when you cannot begin or do not recognise the required process. Enter at FADE when you understand a displayed route but cannot reproduce all of it. Enter at SOLVE when you can explain the process before seeing options or feedback. Use RETURN when the skill appears secure in the current session.

This diagnostic use of a question differs from performing a long unsupported block. The pretesting guide explains how an initial attempt can focus subsequent study even when the first answer is wrong.

How many worked examples should you use?

There is no validated MRCPCH number. Use a performance-based stopping rule:

• study another full example if you cannot identify or reconstruct the process;
• fade support when the route can be explained accurately;
• switch to independent questions when the faded steps are produced without cues; and
• stop using near-identical examples when they no longer change errors or explanations.

A fixed "three examples per topic" rule ignores prior knowledge and task complexity. One clear example may be enough for a familiar process. Several varied examples may be needed when the same principle appears in different question structures.

Count transitions, not pages: example studied, support removed, independent question solved, delayed transfer checked.

How to read an answer explanation actively

Use a four-column note only for explanations that reveal a genuine gap:

Prompt	What to record
Task	What was the question actually asking?
Decision	What information separated the best option?
Error	What did I know, misread, confuse or fail to retrieve?
Transfer	What different question would require the same process?

Then close the explanation and say the route in one or two sentences. If the note cannot be understood without reopening the original question, it is probably a summary of the page rather than a reusable decision rule.

The self-explanation guide gives prompts for making reasoning explicit without producing an essay after every item.

When worked examples become counterproductive

Reduce or remove them when:

• you predict every step before it appears;
• you recognise the wording rather than reconstructing the process;
• explanations feel fluent but an unseen question remains difficult;
• you spend longer polishing notes than attempting transfer;
• the model includes information you no longer need; or
• mixed questions reveal that you cannot select the process independently.

This is not a ban on explanations. Even experienced candidates need worked feedback after a novel error. The issue is unnecessary assistance on a task that should now be retrieved.

When independent questions are being used too early

Add support when:

• several errors share the same missing process;
• feedback is read but the route cannot be reconstructed;
• answers are chosen by option familiarity alone;
• repeated attempts produce the same guess; or
• the candidate cannot describe what would count as improvement.

More questions do not automatically solve an absent model. Pause the block, find one reliable explanation, process it actively and retry with fading.

The question-bank study method shows how to combine attempts, feedback, error diagnosis and reattempts in a complete learning loop.

A practical 45-minute example-to-question session

The timings below are an editable example, not an evidence-based prescription:

1. Five minutes: attempt one representative question and locate the failure point.
2. Ten minutes: study one worked solution, predicting and explaining each important step.
3. Five minutes: hide the solution and reconstruct the route.
4. Ten minutes: complete one faded or partially worked item.
5. Ten minutes: solve one matched but unseen question independently.
6. Five minutes: record the remaining gap and schedule a delayed transfer question.

If the initial question is solved with sound reasoning, skip the full example and move to a less similar independent item. If the worked route still cannot be explained, do not rush into a large question block merely to meet a quota.

Use official RCPCH material for the right purpose

The current RCPCH theory hub says that FOP, TAS and AKP are the three theory examinations and links the current syllabi, question-type guidance and online sample papers. RCPCH describes the online samples as a way to understand the computer-based examination experience.

Use the official material to:

• confirm current scope and terminology;
• see the published question presentation;
• practise the computer experience; and
• anchor preparation to current RCPCH information.

Do not infer that a limited sample predicts your score, represents every syllabus area or supplies enough repeated practice by itself. Use independently authored questions for broader learning, while checking time-sensitive examination claims against the live RCPCH site.

Final answer

Worked examples are most useful when you lack a reliable route through a task. Practice questions are essential when you need to perform that route without support. The efficient sequence is not "read everything, then test" or "answer hundreds and hope". It is SHOW - FADE - SOLVE - RETURN.

Begin with enough assistance to understand the process, actively explain it, remove support, solve a new question and check transfer after a delay. Let demonstrated performance determine the next level of help.

Frequently asked questions

Should I read the explanation before answering an MRCPCH question?

Usually attempt the question first when you already understand the task and can begin a reasoned solution. Read a worked explanation first when the process is genuinely unfamiliar or repeated unsupported attempts are producing guesses rather than learning. After studying it, hide the solution and answer a matched new question independently.

Are worked examples better than practice questions for MRCPCH revision?

Neither is universally better. Worked examples can support initial understanding or repair a missing process; independent questions test whether that process is available without cues. Use more support for an unfamiliar task, fade it as performance improves, and make unsupported retrieval and mixed transfer the destination.

How many worked examples should I study before doing MRCPCH questions?

There is no evidence-based MRCPCH number. Study examples until you can predict and explain the important steps, then move to a partially completed item and an unseen question. If independent reasoning fails, restore the smallest useful amount of support. If examples have become predictable, further near-identical reading is unlikely to be the best next task.

Can I use answer explanations as worked examples?

Yes, if the explanation makes the decision process visible: what the lead-in asks, which information matters, why the best option fits and why close alternatives do not. A statement of the correct answer alone is weak as a worked example. Process the explanation actively, cover it and reconstruct the route before moving on.

When should I stop using worked examples?

Reduce them when you can explain the process without the page, complete faded steps accurately and solve a varied unseen question with sound reasoning. Return to a worked example after a new or repeated process error, but avoid keeping full support merely because the explanation feels fluent.

Sources

1. RCPCH - Theory examinations
2. RCPCH - Theory exams: structure and syllabi
3. A meta-analysis of the worked examples effect on mathematics performance
4. A cornerstone of adaptivity: a meta-analysis of the expertise reversal effect
5. Self-explanation in learning clinical reasoning: the added value of examples and prompts
6. Comparative effects of test-enhanced learning and self-explanation on long-term retention