Final Deadline for Winter 2026 Programme is 21 November. Apply Now →
HomeOCAR AcademyOxford Scholar ProgrammeOxford Publication FellowshipOCAR Student Research SymposiumOxford Essay PrizeOur AlumniOutreachPodcastsResearch FellowsBlogAboutFAQsFor Parents & CounselorsContactAdmissions Portal
A student in conversation with two academics across a table
Admissions Guidance

How to Discuss Research in a University Interview

In short

An interviewer is not checking whether your project was impressive. They are checking whether you understand it, chose it for reasons, and can change your mind when given a better argument. So prepare to justify decisions rather than narrate findings, know your project's weaknesses better than they do, and never defend a claim your evidence cannot support.

If you have a research project in your personal statement, you will probably be asked about it. Most candidates prepare the wrong thing: a polished account of what they found.

What the interview is actually for is different.

What they are assessing

Oxford's computer science department states the general standard plainly: tutors decide "based on your academic abilities and potential alone". On the work you have chosen to do beyond the syllabus, the same page says it would "rather hear about what you've chosen to do, and what excited you about it" — and, importantly, that it "doesn't have to be earth-shatteringly original."

Lady Margaret Hall's description of what Oxford tutors look for names the specific capacity being tested: students who can "present and defend their ideas well but also be open to new insights".

Read that last phrase carefully, because it contains the whole of the advice. Both halves are being marked. Defend your ideas — so you need reasons, not just results. Be open to new insights — so you need to be able to move when someone gives you a better argument.

A candidate who defends every point to the death fails the second half. A candidate who abandons every point at the first push fails the first. The thing being assessed is the judgement about which to do when.

The four questions you will get

Almost every conversation about a student project runs through these, in roughly this order.

QuestionWhat it is really testingLength of answer
"Tell me about your project"Whether you can summarise2–3 sentences
"Why did you choose that method?"Whether you made decisions or followed steps1–2 minutes
"What would you do differently?"Whether you know your own weaknesses1 specific change
"What does this mean?"Whether your claims match your evidenceNarrow, not grand

1. "Tell me about your project."

Answer in two or three sentences, then stop.

The question is an invitation, not a request for a presentation. A candidate who talks for five minutes has chosen the topics themselves and prevented the conversation. A candidate who gives a tight summary hands the interviewer the choice of where to go deeper — which is what they want, and which usually means you end up discussing the part you know best.

Assume they have not read your paper.

2. "Why did you choose that method?"

The highest-value question, and the one most candidates are least ready for.

Narrating the procedure is not an answer. I ran a survey of 84 students describes what happened. The answer they want is why a survey rather than interviews, why 84, why those year groups — and what the design could not show.

This is the same demand academic assessment makes. The AP Research rubric awards 7 of its 50 points for a rationale for the chosen method in relation to the research question, and its top band requires a replicable method. An interviewer is asking the same thing conversationally.

Practise by having someone ask you "why?" four times in a row about one methodological choice. If the fourth answer is "because that's what I did," you are not ready.

3. "What would you do differently?"

A gift, and candidates routinely waste it by being modest in a vague way. I'd have liked a bigger sample is the answer everyone gives and it demonstrates nothing.

A good answer names a specific change and says what it would have bought: I'd have piloted the questionnaire on ten students first, because two items turned out to be ambiguous and I had to drop them, which cost me the comparison I most wanted to make.

That answer shows you know where your study is weak, which is the surest sign you understand it.

4. "What does this mean? / So what?"

The implications question. Who should do something differently, and what?

Also where candidates overreach, and overreaching is the single most expensive mistake available. A measured effect in one school is not a claim about students in general. A correlation is not a cause. Narrow the claim and say what would be needed to support the bigger one — our guide to whether your research is good enough to publish covers the mismatch test, and it applies verbatim at interview.

Know your weaknesses better than they do

The counter-intuitive core of preparing for this.

Interviewers find the flaw in a student project quickly — they do this professionally. You have two options when they do: be surprised, or have got there first.

So before the interview, write down:

That last item is the most useful thing you can bring. A candidate who says my first analysis was wrong because I'd miscoded two variables, I found it when the numbers looked implausible, and here's what I changed is demonstrating research competence more convincingly than any finding.

Three mistakes

Overclaiming. Covered above, and worth repeating because it is the common one. Interviewers test claims by pushing on them, and a claim you cannot support collapses in two questions — taking your credibility on everything else with it.

Rehearsing. A memorised answer is audible, and it prevents the conversation. Prepare the material — your reasons, your weaknesses, your numbers — not the sentences.

Misrepresenting your contribution. If a mentor designed the study or ran the analysis, say so. You will be exposed within two follow-up questions, and the honest version is not a weakness: my supervisor suggested the regression approach and I implemented and interpreted it is a perfectly good answer.

That honesty is also the right standard elsewhere. The ICMJE authorship criteria require an author to be accountable for all aspects of the work, including resolving questions about its accuracy — which you cannot be for an analysis you do not understand. Our guide to student research paper authorship covers the division of credit.

If your project is small

Most candidates' projects are small. This is not the problem it feels like.

A survey of sixty students, analysed properly, understood completely, with its limitations stated precisely, interviews better than a lab placement where you followed someone else's protocol and cannot explain why it was that protocol.

Interviewers probe for depth because depth is what distinguishes genuine engagement from a line on an application. Scale is not available to a seventeen-year-old; understanding is. Our guide to what counts as original research covers why the bar is lower than students assume.

For medicine applicants

One course-specific note. If you are applying to medicine and expecting to be asked whether research or work experience matters more, the Medical Schools Council's own guidance reframes it: on work experience the focus is "on what you learn and how you can reflect on that experience rather than the setting itself", shadowing is "helpful" but "not essential", and it need not be clinical.

The same standard applies to a research project. What you reflected on beats where you were. Our guide to research versus work experience for medicine goes further into it.

The night before

Re-read your own paper. Candidates forget their own numbers under pressure, and nothing undermines you faster than being unsure of your own sample size.

Then write, on one side of paper: your question in one sentence, your method in one sentence, your main finding with its number, your three weakest points, and your one specific "differently".

That page is your preparation. Everything else is the conversation, which is the part you cannot script — and the part they are actually interested in.

Our broader guide to preparing for an Oxford or Cambridge interview covers the rest of the process, and writing research into a UCAS personal statement covers getting it onto the page in the first place.

Frequently asked questions

Will they ask about my research project at interview?

If it is in your personal statement, assume yes. A project is the most interrogable thing on an application — it has choices in it, which makes it ideal material for the kind of conversation interviews are built around. Treat anything you have written down as fair game.

What if my project was small or unfinished?

That is fine and often better. A modest project you understand completely beats an ambitious one you cannot explain. Interviewers probe depth, not scale, and a candidate who can say precisely why their study could not answer a bigger question is demonstrating exactly the judgement being assessed.

What if I get something wrong in the interview?

Say so and work it through. Oxford colleges describe looking for students who can present and defend their ideas well but also be open to new insights — which means revising your position in response to a good argument is the behaviour being tested, not a failure. Defending an indefensible claim is far worse than conceding one.

Should I bring my paper to the interview?

Only if asked. Assume the interviewer has not read it and be ready to summarise it in ninety seconds. Having a copy to hand for your own reference is sensible; producing a document unprompted and expecting them to read it is not.

How much detail should I go into?

Start short and let them choose the depth. Give a two-or-three-sentence summary, then stop and let them ask. Candidates who deliver a five-minute rehearsed account prevent the conversation the interview exists to have.

What if my mentor did most of the work?

Be honest about the division of labour, because you will be found out within two questions if you are not. Describe exactly what you did and what was done for you. That is also the right standard for authorship — supervision does not by itself earn a byline, and the same honesty applies at interview.