What four named Oxford interviewers say they are listening for, and what actually happens when you go quiet
Book a Free ConsultationThinking out loud is not presentation advice at Oxford and Cambridge. It is the thing being assessed. Oxford tells candidates their “final answer is only part of what they find out”, and Cambridge says its interviewers “are interested in how you work out your answers and how you think”. Of roughly 23,000 Oxford applicants, about 10,000 reach an interview in December, and this is what separates them.
You cannot learn this from a list of questions. Our subject interview packs (£180 each) are written by specialist subject tutors as transcripts of reasoning rather than finished answers, so you can see where a strong candidate hesitates, what they say next and where they change their mind. Nine subjects have a free sample.
See the packs and free samples Or book a free consultationMost interview advice treats reasoning aloud as a technique for appearing engaged. Both universities describe it as the assessment itself, and they are unusually explicit about it.
Oxford, in its guidance for applicants: “Don’t worry if you don’t know an answer immediately — see if you can work your way through it aloud. This lets tutors see how you think and means that your final answer is only part of what they find out.”
Cambridge, in its interview preparation guidance, lists among the things to remember on the day that “the people interviewing you are interested in how you work out your answers and how you think”.
Read those together and the consequence is uncomfortable but useful: a silent candidate who arrives at the right answer has withheld most of the evidence. The answer is the smaller part of what is being collected. Everything below follows from that.
Oxford publishes sample questions attributed to named tutors, each with an explanation of what the question is for. Taken together across four subjects, they describe one criterion rather than four.
| Oxford interviewer | Subject | What they say they are listening for |
|---|---|---|
| Kathryn Scott, Christ Church | Biochemistry | “How the candidate applies their chemical knowledge to a problem they are unlikely to have considered before, how they justify their conclusions and whether they are capable of considering alternative possibilities” |
| Robert Wilkins, St Edmund Hall | Biomedical Sciences | “How students approach a clinically-relevant problem” |
| Andrew King, Exeter College | Medicine | A question that “can be approached in many different ways, without necessarily knowing anything about specific mortality rates around the world” |
| Owen Lewis, Brasenose College | Biology | “Once students have come up with a plausible theory, I’d follow up by asking them how they would go about testing their idea” |
Notice how little of that is about knowledge. Scott says outright that “getting the correct answer isn’t important here”. King’s question is designed to be answerable “without necessarily knowing anything about” the underlying figures. Wilkins’s builds on school biology and chemistry to see how a candidate “approaches” something clinical. Lewis is waiting for a theory so that he can ask how you would test it.
Four tutors, four subjects, one thing being measured: what you do with what you already know when the problem is unfamiliar.
The reason candidates stop reasoning aloud is fear of the silence that follows a wrong turn. It is worth knowing exactly what an interviewer does about that, because Oxford describes it.
Kathryn Scott, on her Biochemistry question: “If a student struggles with a starting point we would prompt them to describe more generally the different kinds of interactions that hold molecules together and to comment on their relative strength. We would then encourage them to think about what interactions the specific compound might make first with octanol and then with water.”
Two things follow. First, the prompt widens the question rather than supplying the answer — it is scaffolding, and accepting it is how the exercise is meant to run. Second, and less obviously, a prompt can only be aimed at reasoning the interviewer can hear. If you go quiet, there is nothing to aim at, and a tutor who cannot find your thread has to either wait or move on. Neither helps you.
Cambridge adds a permission most candidates never use: “if you don’t understand something you can ask your interviewers to explain it to you.” Oxford says the same about the question itself — “don’t be afraid to ask for them to be repeated.” Asking is not a confession.
Reasoning aloud under interruption is a skill, and it is very hard to rehearse alone. Our specialist tutors run mock interviews that prompt the way Oxford describes — generalising the problem when you stall rather than moving on — so the first time it happens is not in December.
Book a Free Consultation Message us on WhatsAppOxford gives a method for one of its own questions, and although it was written about an organic compound it generalises further than that: “A good approach to answering the question would be to first consider the individual functional groups separately and then to discuss the compound as a whole.”
Parts, then whole. Applied to almost any unfamiliar problem it produces a usable opening: name the components you recognise, say what each one does on its own, then say what happens when they interact. It has the additional virtue of being audible — each step is a sentence, so the interviewer can follow you and intervene at the point where you go wrong rather than at the end.
Then there is the part candidates leave out. Oxford notes that its published question has an extension, and that “the extension to this question asks the student to interpret some graphical data and requires a more technical introduction.” Owen Lewis describes the same move in Biology: theory first, then “how they would go about testing their idea. What sort of data would they need?”
The second question is usually where the discrimination happens, because most shortlisted candidates can reach a plausible first answer. Volunteering the test — “if that were right, I’d expect to see X, and I could check it by Y” — answers the follow-up before it arrives and demonstrates the habit the follow-up exists to find.
My son was low in confidence after COVID and the gaps in learning were building up. Using this service helped plug those gaps, rebuild confidence, and improve his ability to learn properly again.
We arranged maths and English tuition ahead of GCSEs and both tutors were brilliant. My daughter's grades improved, and by results day she achieved two grades higher than in her mock exams.
I attended an English bootcamp with Leading Tuition and the experience was excellent. The teacher gave focused attention to the small group while keeping the lessons interactive and engaging.
The failure modes are specific, and each one has a university sentence pointing the other way.
| What candidates do | What it costs | What to do instead |
|---|---|---|
| Go silent while thinking | The interviewer cannot prompt reasoning they cannot hear | Narrate the first move, even if it is “I want to start with what holds these molecules together” |
| Guess what the tutor wants to hear | Cambridge says explicitly that you do not need to | Answer in your own way and say why you chose that route |
| Answer a question you did not fully hear | You spend the answer on the wrong problem | Oxford says do not be afraid to ask for a question to be repeated |
| Stop at the first plausible theory | The extension is usually where the discrimination happens | Offer how you would test it before being asked |
The second row is the one worth dwelling on, because it is so widely encouraged elsewhere. Cambridge states plainly: “you can answer the questions in your own way, you don’t need to say what you think the interviewers want to hear.” A candidate performing an imagined preference is not showing their reasoning, they are showing their model of the tutor — which is the one thing nobody is marking.
Both universities give practice advice for 2027 entry, and both are more modest than the industry that has grown up around this.
| The December 2026 round, for 2027 entry | Figure or date |
|---|---|
| Oxford applicant pool | around 23,000 |
| Oxford applicants normally interviewed | around 10,000 |
| Oxford interview format | Online, at least one interview |
| Oxford interview month | December 2026 |
| Cambridge main interview period | 7–18 December 2026 |
| Cambridge winter pool round | Mid to late January 2027 |
Two numbers in that table set the stakes. Around 10,000 of roughly 23,000 Oxford applicants are interviewed, so more than four in ten reach this stage — and once there, the interview is doing most of the remaining work. Oxford puts it plainly: “If you are invited for interview, then you are a strong contender for a place.”
Cambridge: “Have a practice interview. Ideally ask a teacher or someone who doesn’t know you well so that you can practise expressing your ideas and opinions. Alternatively, you could ask yourself questions for half an hour.” The phrase to notice is someone who doesn’t know you well — a parent who can already follow your shorthand will not notice the gaps a stranger hears.
Oxford is blunter still: practice conversations “don’t have to be formal ‘mock interviews’ — instead they could be chats with teachers, friends, or family.” It also settles a persistent myth: if you are offered the chance to ask a question at the end, “this does not form any part of your assessment.”
One constraint shapes what material you can practise on. Oxford’s code of conduct, which candidates accept along with the invitation, includes an undertaking “not to share in any form the content of interviews with others” — so lists claiming to hold this year’s real questions are either old, invented, or a broken promise. What is legitimately available is what the universities publish and what tutors have chosen to make public. Our biochemistry interview guide and physiology guide both work a published Oxford question through in full, and the online interview setup guide covers the tier system that decides whether you will be reasoning aloud on a shared whiteboard as well.
Interviews run online at Oxford in December 2026, and at Cambridge in a main period of 7–18 December 2026. That leaves the autumn to make speaking your working the default rather than the thing you remember to do. Across the students we support through selective applications, 91% go on to achieve the grades they were aiming for — but the interview is a separate skill from the grades, and it is the one that rewards rehearsal most directly.
Cambridge suggests practising with someone who does not know you well. If nobody suitable is to hand, our specialist tutors do exactly that job, in your subject, using published questions — and our subject packs (£180 each) give you the worked reasoning to study between sessions.
Browse the subject packs Or book a free consultationBecause it is what is being assessed. Oxford tells candidates: “Don’t worry if you don’t know an answer immediately — see if you can work your way through it aloud. This lets tutors see how you think and means that your final answer is only part of what they find out.” Cambridge says the same from the other side: “the people interviewing you are interested in how you work out your answers and how you think.” A silent candidate with a correct answer has withheld the evidence.
They prompt. Kathryn Scott, who interviews for Biochemistry at Christ Church, describes it: “If a student struggles with a starting point we would prompt them to describe more generally the different kinds of interactions that hold molecules together and to comment on their relative strength.” The prompt widens the question rather than giving the answer, and it is designed into the process. Being prompted is not a mark against you; it is the conversation working.
Not in the way candidates fear. Oxford states of its published Biochemistry question that “we do not expect the student to know the answer to this question straight away, and in fact getting the correct answer isn’t important here.” What is assessed is how you apply what you know to something unfamiliar, how you justify your conclusions, and whether you can consider alternatives. A wrong answer reached by visible, sound reasoning is worth more than a right one produced from nowhere.
Start with the parts. Oxford sets out the method for one of its own questions: “first consider the individual functional groups separately and then to discuss the compound as a whole.” That generalises well beyond chemistry — break the thing into components you do recognise, say what each one does, then reassemble. Saying “I don’t know, but here is what I would try first” is a legitimate opening and gives the interviewer something to work with.
No, and Cambridge says so directly: “you can answer the questions in your own way, you don’t need to say what you think the interviewers want to hear.” It adds that you can talk freely about your views and share new ideas, and that if you do not understand something you can ask the interviewers to explain it. Second-guessing costs you the thing being marked, which is your own reasoning rather than your impression of theirs.
Cambridge gives two options. The better one: “Have a practice interview. Ideally ask a teacher or someone who doesn’t know you well so that you can practise expressing your ideas and opinions.” The fallback: “Alternatively, you could ask yourself questions for half an hour.” Oxford adds that practice conversations “don’t have to be formal ‘mock interviews’” and can be chats with teachers, friends or family. The requirement is that you speak, not that the setting is formal.
Treating a pause as failure and filling it with silence rather than with reasoning. The second is stopping at the first plausible answer. Owen Lewis, who interviews for Biology at Brasenose, describes what follows a good theory: “Once students have come up with a plausible theory, I’d follow up by asking them how they would go about testing their idea. What sort of data would they need?” A candidate who volunteers that step has already answered the follow-up.
Our specialist tutors run mock interviews that behave like the real thing: an unfamiliar problem built on your syllabus, prompts rather than rescues when you stall, and a follow-up that pushes past your first answer. We work from published questions, since Oxford’s code of conduct means genuine recent questions are not in circulation. Our subject packs are £180 each and set the reasoning out as transcripts rather than model essays; nine subjects have a free sample.
Free samples in nine subjects show you the format, the full expert packs are £180 each and set the reasoning out as transcripts, and 1-to-1 mock interviews put it under real conditions.
See the packs and free samplesLeading Tuition is rated 4.8 out of 5 on Trustpilot, based on 58 reviews from families across the UK. 91% of our students achieve their desired grades.