Revision systems

EPQ ideas: turn an interest into a researchable question

Five worked examples show why a question fails, how to revise it and when to choose a different project.

Editorial

Lead authorRobin GibbonsFounder & CEO·Latimer Tuition
Reviewed byHeather LatimerTutor Matching & Customer Service·Latimer Tuition

Note

From an interest to a defensible project

“Local history” is a topic. A question specifies what you want to find out, which material you will examine and what would count as an answer. For a research review, aim for analysis rather than a catalogue of facts: different interpretations should be possible, but you do not have to give weak evidence equal weight.

OCR’s planning guide supports developing a focused question that allows analysis and evaluation. For a practical outcome, identify what you will make and how you will evaluate it, rather than forcing every project into an essay title.

Qualification details here concern England. AQA and OCR have their own requirements: AQA’s specification includes suitability, manageable scope, resources and genuine extension beyond your programme of study. Discuss overlap with existing coursework before proceeding; the same assessed work cannot simply be submitted twice. Your centre must approve the project under the requirements that apply to you.

Note

Make a shortlist, then run a source trial

Write down three interests and what puzzles you about each. Add the skills you already have and what you would need to learn. OCR’s planning guide recommends exploring ideas and auditing interests and skills before committing.
For each interest, draft your own question and state its boundary: a particular text, place, period, group or design task. Explain why that boundary matters rather than adding a date just to make the title look precise.
Open some relevant evidence. Can you read the full material, understand its method and explain what it can and cannot establish? An abstract or search-result headline is not a substitute for access to the source.
Look for a serious alternative explanation or a trade-off. If you already know what every paragraph will conclude, ask whether the wording has ruled out disagreement.
Record the question, sources tested, remaining problems and your keep/revise/reject decision. Take that record to your supervisor.

Keep source titles, authors, links, search terms and useful page numbers as you go. OCR’s researching and referencing guide discusses research routes such as libraries and academic archives. There is no source quota in this exercise: the test is whether the evidence can answer your particular question.

Note

Example 1 — renewable energy: reject an unworkable scale

Broad topic: renewable energy.

Rejected question: “Can I build a grid-scale wind turbine to power my town?”

Why reject it: the construction, equipment, permissions and scale are not a manageable student investigation. Enthusiasm does not solve that feasibility problem. Making a small model would also change the question: it would not establish whether you could power a town.

Revised teaching example: “How do published comparisons of rooftop solar and small wind assess their suitability for a specified domestic setting?”

This changes the task from building infrastructure to evaluating published evidence. Before developing their own version, a student would have to choose a setting and comparison criteria, such as output assumptions and installation constraints. They would check whether the sources use comparable conditions and explain where they do not. One source might favour output while another emphasises practical siting; that is a comparison to investigate, not a winner to announce in advance.

Decision after the source trial: revise again or reject the idea if usable comparisons cannot be found. A narrower title is not automatically a feasible project.

Note

Example 2 — local history: change the claim to fit the evidence

Broad topic: a local factory closure.

Rejected question: “What did every resident think about the factory closing?”

Why reject it: newspaper reports cannot recover every resident’s private opinion. A few interviews would not fix that overclaim, and access to interviewees is not assured.

Revised teaching example: “How did two local newspapers frame the factory closure during the six months around the announcement?”

The evidence is now the newspapers’ coverage, not the whole town’s beliefs. First check that both archives are accessible and contain enough relevant material. Compare whose voices appear, which consequences receive attention and whether news reports differ from editorials. The newspapers might offer competing economic and community accounts, or they might largely agree; either result must remain possible.

Decision after the source trial: if one archive has major gaps, change the comparison or choose another event. Do not quietly replace missing evidence with an assertion about what residents must have thought.

Note

Example 3 — literature: remove the predetermined answer

Broad topic: fictional villains.

Rejected question: “Why is the antagonist obviously evil?”

Why reject it: “obviously” settles the judgement before the investigation begins. The likely result is a list of supporting incidents rather than an argument that can respond to counter-evidence.

Revised teaching example: “To what extent does one chosen novel present its antagonist as responsible for their actions?”

A student would select a text they can examine closely and check that the project extends beyond their taught work. They could place passages suggesting deliberate choice beside passages suggesting pressure or circumstance. Who tells the story, and how reliable is that perspective? A critical reading that disagrees with the first interpretation gives something substantive to test against the novel.

Decision after the source trial: retain the question only if the text and accessible criticism allow a sustained investigation. Do not manufacture a balanced conclusion: one interpretation may withstand the evidence better than another.

Note

Example 4 — sleep and memory: change an unsafe method

Broad topic: sleep and memory.

Rejected question: “Will deliberately keeping my classmates awake reduce their test scores?”

Why reject it: do not deprive classmates of sleep or recruit them into an informal experiment as a shortcut to original data. A question can be interesting while its proposed method is inappropriate.

Revised teaching example: “How do published studies differ in their explanations of the relationship between sleep and memory in a defined age group?”

The task becomes a review of existing research, subject to supervisor agreement. A student would select an age group after checking what evidence is accessible, compare what the studies actually measured and distinguish an observed association from a claim about cause. If studies use different memory tasks or populations, that limits the comparison; it is not a detail to erase.

Decision after the source trial: take any safety or sensitive-topic concern to the supervisor before collecting data. AQA’s safeguarding guidance requires careful assessment of topics and primary research, including sensitive topics with a personal connection. Published evidence is a different method, not automatic approval for any topic.

Note

Example 5 — accessible design: keep the core work yours

Broad topic: accessible app design.

Rejected question: “Can my parent design and code my app while I document it?”

Why reject it: the central design decisions and creation would belong to someone else. A polished product would not demonstrate your own project work.

Revised teaching example: “How could I design and evaluate a paper prototype for one navigation task against published accessibility criteria?”

A limited prototype could make the student’s own reasoning visible without depending on someone else’s programming. They would choose the task, identify applicable criteria, create alternatives and explain why they retained one design. For example, reducing the number of screens might conflict with keeping each screen easy to understand; the evaluation should explain that trade-off rather than simply calling the design user-friendly.

Decision after the source trial: establish whether the selected criteria can genuinely be assessed on paper. A paper prototype cannot prove that a finished app is accessible to all users. Discuss any proposed user testing and arrangements with the supervisor first.

Note

Choose between two promising ideas

Compare evidence access, scope, skills, safety and interest separately. An exciting question with inaccessible essential evidence is not rescued by scoring highly for enthusiasm.

Suppose your history idea has two readable archives but your energy idea depends on an unavailable technical dataset. Keep the history idea provisionally; revise the energy question so that it uses evidence you can obtain, or set it aside. If both pass, ask which uncertainty you most want to resolve and which method you can learn within your available time. This is a decision aid, not an assessment mark scheme.

Write a short proposal in your own words: what you want to investigate or create, why the boundary is sensible, which evidence you have tested, what remains uncertain and your next step. Keep the rejected version and explain the change. AQA’s production log records development and reasons for changes, so a justified revision is part of the project rather than something to conceal.

Related Ed Centre pages

These linked pages help students and parents move between closely related guidance instead of reaching a dead end.

Section overview

Revision systems and study planning guides

Find published student guides on revision methods, planning systems, timetables, active recall, flashcards and realistic catch-up routines.

Related guide

How to revise A-level Biology: practicals and essays

Start with your exam board, then train the practical, data, command-word and synoptic-writing skills that written papers actually test.

Related guide

A-level Maths revision plan by paper and topic

Build your plan from your exam board first, then audit Pure, Statistics and Mechanics, repair weak topics and move into timed mixed-paper practice.

Related guide

Active recall and spaced repetition: a simple revision system

Learn how to test yourself from memory, space topics out and keep your revision system simple enough to use.

Related guide

Blurting revision method: a worked correction cycle

See an incomplete GCSE answer, diagnose what is missing or wrong, correct it and test the learning later.

Related guide

How to Build a Revision Error Log That Turns Mistakes Into Useful Revision

Diagnose why an answer went wrong, record the right fix and retry the skill later without the solution in front of you.

Related guide

Cornell note taking: turn a page of notes into a self-test

Copy the layout, follow a worked GCSE example and see how checking your answers changes the summary.

Related guide

EPQ presentation examples: build your own explanation

See what belongs on a slide, what to explain aloud and how to respond when someone challenges your conclusion.

Related guide

The Feynman technique: explain, check, then apply

Follow a flawed circuit explanation through a precise correction and a fresh question—then try the process on your own topic.

Support and clarity

Frequently asked questions

Straight answers to the questions people ask most often.

Can my EPQ question change after I start reading?

Yes, a working question can develop as you discover limits or better evidence. Discuss a significant change with your supervisor and record what changed and why. For AQA, the production log provides a record of that development; do not rewrite the history as though the final question existed from day one.

What if most of the useful articles are behind paywalls?

Ask your library about legitimate access routes and check academic archives. If you still cannot access evidence central to the question, revise the scope or choose another idea. Do not build the argument from search snippets or assume an abstract supports details you have not read.

Does an EPQ question have to begin “To what extent”?

No particular opening makes a question analytical. Ask whether it identifies a manageable investigation, leaves room for evidence to affect the answer and suits the intended outcome. A practical design project also needs a clear way to evaluate what you make, not just an essay-style title.

1.
AQA EPQ specification v1.4
AQA
2.
Extended Project Qualification safeguarding requirements
AQA
3.
Guidance and support for the use of AI in the Extended Project Qualification
Cambridge OCR
4.
Guide 1: Planning, version 4.0 March 2026
Cambridge OCR
5.
Guide 2: Researching and referencing, March 2015
OCR

Continue exploring Revision systems and study planning guides

Frequently asked questions