Put the project in Question 3, in roughly 700–900 of your 4,000 characters, and write about the part that went wrong. Oxford's own guidance is that "reeling off a list of books you have read is not helpful" — the same applies to reeling off a project. Tutors are reading for evidence you did the thinking, not proof you completed a thing.
Most students write about their research project twice as confidently as they should, and half as specifically.
The instinct is to present it as an achievement: a title, a method, a result, ideally a publication. That instinct is wrong, and it is wrong in a way that is visible to anyone who reads applications for a living. The polished version reads like a press release. The version that works reads like someone describing a problem they actually had.
This article covers where the project goes, how to write it, what to do if it is unfinished, and the rules on AI that now apply to everything you submit.
Which of the three UCAS questions does research belong in?
Question 3, almost always.
Since 2026 entry, the personal statement is three questions sharing 4,000 characters, each with a 350-character minimum:
| Question | What it covers | Where research fits |
|---|---|---|
| 1. Why do you want to study this course or subject? | Motivation and specific interest | A clause, if the project is why you chose the subject |
| 2. How have your qualifications and studies helped you to prepare? | Your formal schoolwork | Only if the project ran through school — an EPQ or IB Extended Essay |
| 3. What else have you done to prepare outside of education, and why are these experiences useful? | Everything self-directed | Here. This is what the question is asking for |
The one genuine exception is the EPQ or Extended Essay. Those sit inside your formal education, so they belong in Question 2 — though nobody will penalise you for placing them in Question 3, and Cambridge has said via Girton College that it does "not recommend any particular split to the three sections."
What you should not do is spread one project across all three. Repetition is expensive when you have 4,000 characters and three questions to fill.
How much space should it take?
Between 700 and 900 characters — roughly 120 to 150 words — is right for most applicants.
That feels short. It is short. It is also more than Oxford's guidance implies you should spend on any single item, given that they recommend around 80% of the whole statement go on academic interests, abilities and achievements collectively.
The constraint is useful. At 900 characters you cannot narrate the project chronologically, which is what most first drafts do. You are forced to choose the two or three things that actually demonstrate something, and choosing is the skill being assessed.
What should the paragraph actually contain?
Four things. In this order, if you have no better idea.
The question, stated concretely. Not the field, the question. "I investigated economic inequality" tells a tutor nothing. "I tested whether minimum wage rises in three UK regions since 2015 tracked local employment changes" tells them you had a testable question and knew what data would answer it.
The problem you hit. This is the beat almost everyone omits, and it is worth more than the rest of the paragraph combined. Research that proceeded smoothly is research nobody learned from — and, more practically, a tutor who reads about a difficulty knows you were present for it. Nobody invents a boundary-definition problem they did not have.
What it changed in how you think. The "why are these experiences useful" half of the question. Be specific about the transferable thing: a method you now distrust, a debate you understand differently, a question you did not know existed.
Where it points. One clause connecting it to the course. Not a declaration about your future.
Two versions of the same project
Here is the version most students submit first:
"I completed an independent research project on the minimum wage, working with a PhD mentor over five months. This developed my analytical and time-management skills, deepened my understanding of labour economics, and confirmed my passion for the subject."
And here is the same project, written properly:
"Over five months I tested whether minimum wage rises in three UK regions since 2015 tracked local employment changes. My first dataset was unusable — regional boundaries changed in 2019 and I had been comparing different places without realising. Rebuilding it taught me more than the result did: that the hardest part of empirical economics is establishing that your data measures what you think it measures. I had read about identification strategies before. I had not understood why anyone cared until my own numbers fell apart."
Both passages are illustrative rather than a real student's work. The second is 560 characters. It contains a method, a failure, a lesson and an intellectual position, and any interviewer could get fifteen minutes out of it.
The first mentions a PhD mentor. The second does not, and does not need to — what the mentor contributed is visible in the quality of the thinking, which is the only place it should be visible.
What if the project is not finished?
Write about it anyway, in the present tense, and say where it has got to.
"I am currently analysing…" and "the paper is under review at…" are both honest, both normal, and both give an interviewer something live to ask about. An unfinished project you can discuss beats a finished one you cannot.
This matters more than students expect, because the timelines rarely align. A project begun in the summer of Year 12 will often still be in review when the 15 October Oxbridge deadline passes. That is the ordinary case, not a failure. The application timeline is unforgiving, and no admissions tutor expects a seventeen-year-old to have completed peer review by the autumn of Year 13.
What you must not do is round up. "Published in" when you mean "submitted to" is the kind of small inflation that, if noticed, costs you far more than the extra word was worth.
What are the rules on using AI to write it?
Strict, checked, and worth understanding precisely — particularly if you have been working with an AI tool during your research.
UCAS is explicit that "generating (and then copying, pasting and submitting) all or a large part of your personal statement from an AI tool such as ChatGPT, and presenting it as your own words, could be considered cheating by universities and colleges." When you submit, you declare that your statement has not "been copied or provided from another source, including artificial intelligence software."
UCAS also runs checks "to detect fraudulent applications and patterns of similarity in personal statements," and where something is flagged, "the universities or colleges it is intended for may be notified." The stated consequence is that you "don't receive any offers or have offers withdrawn."
What UCAS does permit is using AI for brainstorming, structuring and checking readability of work you wrote. The line is between shaping your own words and producing words for you.
There is a practical dimension here beyond the rules. A statement generated for you will be fluent, general and free of the specific detail that makes a project credible — which is precisely the failure mode described earlier. The passage about the 2019 boundary change cannot be generated, because only you know it happened. The best defence against sounding machine-written is also the best writing advice: put in the things only you could know.
What makes a tutor doubt you did the work?
Four things, in rough order of how often they appear.
A question too sophisticated for your training. If your school offers no biology past AS-level and your paper claims findings in computational neuroscience, the mismatch does the talking. Modest and genuine beats ambitious and implausible every time.
Prose that does not match the rest of your application. Your personal statement, your teacher's reference and your interview should sound like the same person. A paragraph written in a register you cannot sustain in conversation is a liability the moment you are asked about it.
No difficulties anywhere. Real projects have friction. A description with none reads as a summary of a project rather than an account of doing one.
Methods you cannot explain. This is what interviews test. Oxford's English Faculty, for instance, assesses submitted written work for "Evidence of careful and critical reading" and "Independence of thought," and probes the same qualities at interview. A tutor asking why you chose that sample size settles the question in about two minutes. Our interview preparation guide covers what that conversation is actually like.
The safest position is the true one. If a mentor shaped the project heavily, write about the part that was yours.
Does mentioning a paid programme help or hurt?
Neither, mostly — and there is a better use of the characters.
Naming a programme tells a tutor where you did the work, which is not what they are assessing. It uses perhaps 40 characters that could carry a finding. And competitive admissions offices see the well-known programme names recur across thousands of applications, so the name adds no distinction.
Write about the research. If the programme comes up at interview, say what it was plainly; there is nothing to hide in having had good supervision, and paid tuition is not cheating. The mistake is letting the brand stand in for the substance.
How does this differ for US applications?
Substantially, and applying to both is where students come unstuck.
The Common App rewards narrative, personal voice, and breadth across activities. UCAS rewards subject depth and nothing else — Oxford states that tutors "are only interested in your academic ability and potential," and Christ's College Cambridge that "we only assess you on academic criteria."
Write the UCAS statement first, at full academic density, then adapt outward for the US. Doing it the other way round produces a UCAS statement with a childhood anecdote at the top and 600 characters of your subject at the bottom, which is the single most common shape of a weak international application to a UK university. We cover the wider differences in our guide to UK versus US applications.
Frequently asked questions
Should I name the journal my paper was submitted to?
Only if it is doing work for you. Naming a selective venue is informative; naming one nobody has heard of is not, and invites a question about its standards. "Submitted to a peer-reviewed journal" is often the better use of the space, with the details ready for interview.
Can I mention my mentor by name?
There is little reason to. Tutors are assessing you, not your supervisor's institution, and a name can read as an appeal to borrowed authority. If the supervision was substantial, the right place for it is your mentor's reference, where it carries far more weight.
How do I write about research if the results were negative or inconclusive?
Exactly as you would otherwise, and with more confidence than you feel. A null result you can explain demonstrates more understanding than a positive one you cannot. Say what you expected, what you found, and what you now think the explanation is.
Is 900 characters really enough for six months of work?
Yes, and the compression is part of what is being assessed. Academics write 250-word abstracts for five-year projects. If you cannot say what mattered in 150 words, that is usually a sign you have not yet decided what mattered.
Will admissions tutors think a paid research programme is unfair?
There is no evidence UK admissions penalise supervised work, and school-supported projects like the EPQ are supervised too. What draws scepticism is a project that reads as though someone else did it. Write it so that yours plainly does not.