The EPQ's defining feature is that nothing constrains it — any topic, essay or artefact, no subject tie. That freedom is why it is worth doing and why students stall in February. The fix is to impose a constraint in week one: pick one discipline, pick one dataset or source base, and let the question come out of those.
The EPQ is the most flexible qualification in the UK system and the easiest to get stuck in. Both facts come from the same source: nothing tells you what to do.
What it actually is
A standalone qualification in which you choose, plan, research and deliver your own project. Two routes:
The dissertation route. Around 5,000 words answering a question you set.
The artefact route. You make something — software, a composition, a designed object, an event, a performance — with a shorter written report alongside.
Both carry a production log recording decisions as you make them, and both end in a presentation. The DOAJ's guidance on evaluating a venue is worth reading before you plan a project around publishing it. The log and presentation are assessed, which surprises students who treat them as paperwork.
AQA allocates 20% of the available marks to AO2: "research, critically select, organise and use information." The phrase doing the work is critically select. You are assessed on adjudicating between sources that disagree — so a question answerable from one textbook cannot score well, however neatly it is written.
The artefact route, which almost nobody takes
Worth a paragraph of its own, because most students do not know it exists.
If you would rather build than write, you can. A piece of software that does something real. An original composition with an analytical report. A designed object with a documented iteration process. An event you planned and ran, evaluated properly.
The research rigour still applies — you still have to show critical selection between sources, still keep the log, still present. But the output can be a thing.
The failure mode here is building something and treating the report as an afterthought. The artefact is not the project; the artefact plus the reasoning behind it is.
Why EPQs stall in February
Because freedom is a cost as well as a benefit, and nobody says so in September.
The IB Extended Essay must sit inside one subject and fit 4,000 words. Those constraints feel limiting and they do enormous work: they force a decision early. The EPQ has neither, so a student can read for four months without ever converging, and by February there is no time to narrow properly.
The fix is to impose the constraint yourself in week one.
Pick one discipline and stay in it. Decide you are doing an economics project, or a history project, and answer accordingly. You will lose some interesting cross-disciplinary options. Lose them.
Pick your source base before your question. Not "I'll research teenage sleep" but "I'll use the Monitoring the Future survey." Narrow questions come from data, not from brainstorming — our list of free datasets covers twenty-two you can open today.
Write the question with its variables named. If it does not contain the specific things being compared, it is still a topic.
Choosing a question that can be finished
One test, applied ruthlessly: can you say in one sentence what evidence would change your answer?
"Is social media harmful to teenagers" fails. No obtainable evidence settles it, and 5,000 words will produce a tour of other people's opinions.
"Does self-reported daily screen time predict self-reported sleep duration among US 12th graders in the 2023 Monitoring the Future survey, controlling for sex and region" passes. The dataset exists, the analysis is reachable, and a result could surprise you.
Two more checks before you commit:
Does someone disagree? Find two sources taking different positions. That disagreement is where your AO2 marks live.
Would any result surprise you? If not, the question is closed and you will spend a term confirming what you already believe.
Thirty worked EPQ questions are grouped by field with a source attached to each, and how to build a question worth publishing is the underlying method.
A timeline that survives exam season
The EPQ's other hazard is that it runs alongside A-levels, and A-levels win every time.
| When | What |
|---|---|
| Weeks 1–2 | Pick discipline and source base. Draft three candidate questions. |
| Weeks 3–4 | Kill two. Write the final question with variables named. Start the log properly. |
| Weeks 5–10 | Read and gather. Write as you go — never read for six weeks and then start. |
| Weeks 11–14 | First full draft. It will be bad. That is what first drafts are. |
| Weeks 15–18 | Revise, cut, resize the conclusion to what the evidence supports. |
| Before exam season | Finish. Everything left undone by then will not get done. |
The single most useful habit is writing from week five. Students who read first and write later find in February that they cannot remember why they saved half their sources — and the production log, which exists precisely to prevent this, is usually the thing they neglected.
What universities make of it
The EPQ has one concrete, published advantage that no other school research project has in the UK: several universities make reduced offers for a strong grade. Check the specific courses you are applying to, because it varies.
Beyond that, its value is the same as any project's — material at interview. For comparison, Caltech is one of the few universities anywhere that publishes a figure, reporting that 42% of admitted students submitted research. A tutor can question you about your EPQ for as long as they like, and a project you genuinely did keeps giving while one that was shaped for you runs out quickly.
Our piece on whether research programmes are worth it in the UK covers what UK universities actually publish about research projects, which is less than the market implies. The short version: if you are weighing paying for a research programme against doing your EPQ properly, do the EPQ properly. It is free, it is on your timetable, and it is the same skill.
EPQ or Extended Essay?
If you are in a position to choose — rare, since the EE is compulsory for IB Diploma students — the differences are set out in full in our comparison of the two.
The short version: the EE gives you discipline and a word cap; the EPQ gives you freedom and an artefact option. The EPQ is harder to start and easier to shape; the EE is harder to scope and easier to begin.
Can you publish it?
Usually not as written, and that is worth knowing in September rather than March.
An EPQ dissertation is an extended argument with a production log. Research journals want original findings and a methods section somebody could repeat. Converting means narrowing the question, adding methods and cutting the reflective material — real work rather than a reformat.
If publication interests you, pick a question with a dataset behind it now. Then check the venue's scope: NHSJS, the Columbia Junior Science Journal and the Oxford Journal of Student Scholarship accept literature reviews, which most student journals do not, so a synthesis-shaped EPQ has somewhere to go. The complete guide to publishing covers the route.
One constraint to plan around early: if your question needs a laboratory, most universities bar under-eighteens from BSL-2, radioactive and animal work. Work in the computational corner of your field rather than waiting for replies that will not come.
Where we fit
The Oxford Centre for Advanced Research pairs students with PhD candidates and early-career researchers at Oxford, Cambridge, other leading UK universities and Ivy League institutions. The Oxford Scholar Programme is £2,000 — about $2,600 — for ten contact hours over ten to fourteen weeks.
The EPQ comes with a supervisor built in. For most students, that supervisor plus a teacher with a doctorate is enough, and the whole method above is free and on this page. Where students come to us is when they want to take the project past what the qualification requires — usually toward submission.
The thing to do in week one
Pick your discipline and your source base before you pick your question, and write both down.
Students who do that finish in the summer. Students who keep their options open are still reading in February, and February is too late to narrow.
Frequently asked questions
What is the EPQ?
A standalone Level 3 qualification in which you plan, research and deliver your own project — either a dissertation of around 5,000 words or an artefact with a shorter written report — alongside a production log and a presentation. It is subject-agnostic, so the topic is entirely yours, and it carries UCAS points in its own right.
Does it have to be an essay?
No, and this is its most underused feature. The artefact route lets you make something — software, a composition, a designed object, a performance, an event — with a written report alongside it. The research and evaluation still have to be rigorous, but the output does not have to be prose.
How much does the research actually count?
AQA allocates 20% of the marks to AO2, defined as 'research, critically select, organise and use information'. The operative phrase is critically select — you are assessed on choosing between sources that disagree, which means a question answerable from a single textbook cannot score well however neatly it is written.
Why do so many EPQs stall?
Because nothing external forces a decision. With no subject tie and no fixed format, students read widely for months without converging on a question. The students who finish make an arbitrary-feeling constraint early — one discipline, one source base — and accept that it closes off options.
Do universities care about the EPQ?
Several UK universities make reduced offers for a strong EPQ grade, which is a concrete published advantage no other school research project has in the UK system. Beyond that, its real value is at interview, where a project you genuinely did gives a tutor twenty minutes of material.
Can an EPQ be published?
Usually not as written — it is an extended argument with a production log, while journals want original findings and a methods section. If publication interests you, choose a question with a dataset behind it from the start and the conversion becomes realistic rather than a rewrite.