The Scholar's Vault · 150 questions, free, with a marker
A hundred and fifty questions, and you only need to find four.
PhD viva and dissertation defence questions, by discipline and by the part of the thesis they attack
This is not a list to memorise. It is a diagnostic: read the eighty that face every thesis plus the seven for your discipline, mark each one, and you will finish with a short list of questions you cannot yet answer while there is still time to do something about it.
No ghostwriting, ever · Human-led, PhD-signed
Every question is on this page, and so is the marker. No address, no account, and nothing you mark leaves your browser. What the email sends is the paper version — a printable with a mark box beside every question, the log sheet as a fillable table, and your discipline’s seven on their own page, because you carry paper into a mock viva and nobody carries a laptop. One email covers all six Vault resources.
How to use it
Eighty-seven questions, about 45 minutes.
Read each question aloud. Do not answer it. Mark it in one of three ways, in under 20 seconds.
| Mark | Meaning | What it costs to fix |
|---|---|---|
| G | You could answer this now, out loud, for 60 seconds | Nothing |
| A | You know the answer and would not say it well under pressure | 20 minutes of rehearsal |
| R | You do not have an answer, or you have one you do not believe | Days, sometimes a re-analysis |
The output. Your R list. In most theses it runs to four to eight questions, and two of them are the same question wearing different clothes.
What to do with it. Work the R list in order of how central the question is to your main claim, not in order of how uncomfortable it is. The two are rarely the same, and the uncomfortable ones are usually amber.
The rule that makes this work: mark honestly on the first pass, before you start reasoning about what an examiner would actually ask. The second pass is always more generous than the first, and the first is the accurate one.
What examiners are testing, section by section
Eight checks, one per part of the thesis.
A viva is not a quiz about your thesis. It is a set of checks on whether the person who wrote it understands what it does and does not establish.
| Section | The underlying test |
|---|---|
| Opening | Can you state the contribution without reading it |
| Literature | Do you know why you are not in a different conversation |
| Theory | Did the framework do work, or was it decoration |
| Method | Did you choose, or did you inherit |
| Analysis | Do you know what your numbers or your codes cannot support |
| Findings | Can you separate result from interpretation |
| Limitations | Do you know where the thesis is weak before they tell you |
| Contribution | Would a stranger in the field be able to use this |
The fear that examiners are hunting for an error you have not noticed is worth naming, because it drives most defence preparation. What they are usually doing is finding the boundary of your claim and checking that you know where it is. That is a checkable thing, which is why a list works.
Answering shape
Three parts, and candidates omit the third.
Most answers are 60 to 90 seconds and have three parts. Say them in this order and the shape does the work.
- The direct answer, first. One sentence. Examiners lose patience with preamble faster than with a wrong answer honestly given.
- The evidence. Where in the thesis, what the result was, what it rests on.
- The boundary. What this does not establish, in one clause.
Part 3 is the one candidates omit, and it is the part that converts a defensive answer into an authoritative one. A candidate who names the limit of their own claim has already answered the follow-up.
For a question you cannot answer: say so, say what you would need in order to answer it, and say what you think the answer probably is. All three. Silence and invention are the only two failing responses.
The format you are walking into
Six rooms, six different fortnights of preparation.
The questions are the same everywhere. The room is not, and candidates rehearse for the wrong room more often than they rehearse the wrong answers.
| Format | Who is in the room | How it runs | What it changes about your preparation |
|---|---|---|---|
| UK and Irish viva | Two or three examiners, one internal, one external, sometimes an independent chair. No audience | Two to four hours, closed, conversational, page-by-page | Every page is fair game. Know your own chapter 4 better than the examiners do, and expect follow-ups rather than set pieces |
| US public defence | Committee of four or five, plus an open audience | Public presentation of 30 to 45 minutes, public questions, then a closed committee session | The talk carries half the assessment. Rehearse it to time, and prepare for the closed session separately |
| Nordic opponent model | A named opponent, a grading committee, a public audience | Public disputation, the opponent leads a long structured critique | The opponent’s line of attack is usually visible in their own published work. Read it |
| Dutch and Belgian defence | A committee of professors, ceremonial, public | Short, timed, formal questions in sequence, often 45 to 60 minutes | The manuscript is usually approved beforehand. Answers are short, and running long is the main failure |
| Australian and New Zealand | Examiners report in writing, and an oral is not always held | Written reports first, revisions specified, oral only in some institutions | Your written response to reports matters more than any oral rehearsal — the reviewer-response letter is the resource for it |
| French soutenance | A jury with named rapporteurs, public | Presentation, then each jury member in turn, often three hours or more | Each jury member has a distinct role. Ask your supervisor who is assigned what before the date |
Ask your supervisor two things this week: which format your institution uses, and who the examiners are. Both change what you should spend the next fortnight on. If yours reports in writing first, the reviewer-response letter is the resource that matters more than any oral rehearsal.
Part 1 · The eighty questions every thesis faces
The bank.
Nine sections, then seventy more in ten discipline sets. Mark as you read: it is stored in your browser, we never see it, and there is nothing to sign up for.
Marking needs JavaScript. Every question is on this page either way, and the printable has a box beside each one.
Section A · The opening two minutes
- Q1
Summarise your thesis in three minutes, as if to an intelligent person outside your field.
- Q2
What is the single most important contribution of this work?
- Q3
What is the one sentence you would want an examiner to write in the report?
- Q4
Why did you choose this topic, and would you choose it again?
- Q5
What has changed in your thinking between the first chapter you wrote and the last?
Section B · Positioning and the literature
- Q6
Which three works is this thesis most in conversation with, and where does it disagree with each?
- Q7
Name the strongest published argument against your position. How do you answer it?
- Q8
Which body of literature did you decide not to engage with, and why was that defensible?
- Q9
Who would be most annoyed by this thesis, and on what grounds?
- Q10
Your review covers [FIELD]. Why not [ADJACENT FIELD], where similar questions are asked with different tools?
- Q11
What is the most recent work that could have changed your design, and when did you find it?
- Q12
Is your gap a gap, or has nobody thought it worth closing? Make the case.
- Q13
Which of your citations are load-bearing, and which are decoration?
- Q14
If the key paper you build on were retracted tomorrow, what survives in your thesis?
Section C · Theory and framework
- Q15
What work does your theoretical framework do that a simpler description would not?
- Q16
What would count as evidence against your framework?
- Q17
You use [CONCEPT] throughout. Define it, and say where your definition departs from the standard one.
- Q18
Where in your thesis does the theory constrain the empirical work, rather than sit beside it?
- Q19
Could your findings be explained equally well by [RIVAL FRAMEWORK]? Why not?
- Q20
Are you testing the theory, applying it, or extending it? Your chapters do not all answer this the same way.
- Q21
Which assumption of your framework is the least defensible, and what happens to your conclusions if it fails?
- Q22
Did the framework come before the data or after it? Answer honestly.
- Q23
What is the relationship between your conceptual model and the model you actually estimated or coded?
Section D · Design and method
- Q24
Why this design, and what was the second-best design you rejected?
- Q25
What would you have done with twice the time and twice the budget, and would the answer be different?
- Q26
Your sample is [DESCRIPTION]. What population does it let you speak about, and what does it not?
- Q27
How did you decide when you had enough data?
- Q28
What did you do about the cases that did not fit?
- Q29
Which methodological choice would you most expect a reviewer to attack, and what is your answer to them?
- Q30
How would somebody replicate this? Walk through what they would need from you.
- Q31
What went wrong during data collection, and what did you change as a result?
- Q32
What ethical judgement did you have to make that the approval process did not decide for you?
- Q33
Where does your method depart from the standard practice in your field, and why was that necessary?
Section E · Data, measurement and analysis
- Q34
What exactly is your unit of analysis, and is it the same in every chapter?
- Q35
How do you know your key measure captures the construct you say it does?
- Q36
What is the effect of your decisions about missing data on the results?
- Q37
Show me the analysis that you ran and did not report. Why is it not in the thesis?
- Q38
Which of your results is most sensitive to an analytical choice, and which choice?
- Q39
What alternative specification, coding frame or model did you test, and what did it show?
- Q40
How much of your finding could be an artefact of how you measured it?
- Q41
What robustness checks would a sceptical reader ask for that are not in here?
- Q42
If somebody re-ran your analysis with a different reasonable set of choices, what is the range of answers they could get?
- Q43
Which of your numbers do you most expect to be wrong in five years?
Section F · Findings and interpretation
- Q44
State your main finding in one sentence, without hedging.
- Q45
Which of your findings surprised you, and what did you do to check it?
- Q46
Where in the thesis do you move from result to interpretation? Show me the sentence.
- Q47
Your conclusion says [CLAIM]. Which specific result supports that, and how directly?
- Q48
What is the most plausible alternative explanation for your central result?
- Q49
Which finding is real but unimportant, and which is uncertain but important?
- Q50
Do any of your chapters disagree with each other? How do you resolve it?
- Q51
What would a null result have looked like in this study, and how close did you come to one?
- Q52
If one of your three findings had to be dropped for publication, which one goes?
Section G · Limitations and validity
- Q53
What is the most serious limitation of this work? Not the one in the limitations section, the real one.
- Q54
What can this thesis not tell us that a reader might think it does?
- Q55
Where are you most likely to be wrong?
- Q56
How far do your findings travel beyond this setting, sample or period?
- Q57
Which limitation could you have designed out, and why did you not?
- Q58
Is there a result here that you believe more strongly than your evidence supports?
- Q59
How would you respond to a reader who said this is a study of [SMALLER THING] presented as a study of [BIGGER THING]?
- Q60
What did you leave out of the thesis, and would an examiner agree with the decision?
- Q61
What in this thesis would you now defend less strongly than you did while writing it?
Section H · Contribution and what comes next
- Q62
What can somebody do now that they could not do before this thesis?
- Q63
Which chapter is the paper, and where would you send it?
- Q64
Who is the audience for this work, and what would you want them to change?
- Q65
What is the next study, and why is it the next one rather than a different one?
- Q66
If you had to abandon this line of research tomorrow, what would you want somebody else to pick up?
- Q67
What did you learn about doing research that is not in the thesis?
- Q68
Is this a contribution to method, to theory, to evidence or to practice? Rank them.
- Q69
What would make this a book rather than three papers?
- Q70
Ten years from now, which part of this thesis do you expect to still be standing?
Section I · Research practice, tools and openness
These are the questions examiners have added most recently. Most candidates have never been asked them out loud, which is exactly why they belong on an R list.
- Q71
Where did you use generative AI in producing this thesis, and how did you verify what it produced?
- Q72
Your institution requires a declaration on the use of AI tools. Talk me through yours, and what you decided not to declare and why.
- Q73
Which parts of this thesis could a model have written, and what does that tell you about the contribution?
- Q74
Was any part of this study preregistered? If not, which analytical decisions were made after you saw the data?
- Q75
Where are your data, your code and your materials, and what would somebody need in order to use them?
- Q76
What in your data cannot be shared, and what is the reason: consent, law, commercial agreement or your own choice?
- Q77
Who has checked your analysis code or your coding frame apart from you?
- Q78
What did your ethics approval not cover that you ended up doing?
- Q79
Which of your decisions would you now register in advance if you started again?
- Q80
What is the version of this thesis that a reader in ten years can still reproduce?
Part 2 · Seventy questions by discipline
Ten sets of seven. Read your own set, then read one other, because examiners from a neighbouring field ask the questions your own field has stopped asking.
Quantitative social science, economics and management
- Q81
What is your identification strategy, and what exactly does it assume about what is unobserved?
- Q82
Your standard errors are clustered at [LEVEL]. Defend that level rather than the one above or below it.
- Q83
How much of this relationship survives if the two most influential observations are removed?
- Q84
Was this hypothesis specified before you saw the data? What is the evidence for your answer?
- Q85
Your effect is statistically distinguishable from zero. Is it distinguishable from trivial?
- Q86
How many models did you estimate before this one, and what would the others have shown?
- Q87
Your construct is measured by [SCALE]. What does it measure in a population that is not the one it was validated on?
Qualitative and interpretive research
- Q88
How did your presence change what your participants said, and where can I see that in the data?
- Q89
Walk me through how you got from a transcript to a code to a theme. Take one theme.
- Q90
Which of your themes is most dependent on a single participant?
- Q91
What did you find that contradicted your developing account, and what did you do with it?
- Q92
What is the claim to knowledge here, if it is not generalisability?
- Q93
Who is not in your sample, and what would they have said?
- Q94
Where in the thesis is your own position doing analytical work rather than being declared and then ignored?
Life sciences and lab-based research
- Q95
What is your negative control, and what would it have shown if the assay had failed?
- Q96
How many biological and how many technical replicates, and why that number?
- Q97
What is the effect size, and is the study powered to detect it?
- Q98
Which of your results depends on a reagent, antibody or cell line whose identity you cannot fully verify?
- Q99
If this mechanism is right, what else must be true, and did you test it?
- Q100
Which experiment in here has been repeated by somebody other than you?
- Q101
What is the single figure that would collapse if the [SPECIFIC ASSUMPTION] were wrong?
Clinical and health professions research
- Q102
What was your primary outcome at registration, and is it the primary outcome in this thesis?
- Q103
Who was excluded from your study, and how does that change who your finding applies to?
- Q104
What is the minimum clinically important difference here, and did you reach it?
- Q105
How would this change what a clinician does on a Tuesday, and what would have to be in place first?
- Q106
What harms did you measure, and what harms could this intervention plausibly cause that you did not measure?
- Q107
How did you handle loss to follow-up, and what does the pattern of dropout suggest?
- Q108
Was patient or public involvement part of the design, and what did it change?
Engineering, computing and modelling
- Q109
What is the baseline you compare against, and is it the strongest available one?
- Q110
Where does your model break, and did you report the conditions under which it does?
- Q111
How sensitive are your results to hyper-parameters, initialisation or mesh choice?
- Q112
What did you validate against reality, as opposed to against another model?
- Q113
What is the computational cost of your method, and at what scale does the advantage disappear?
- Q114
What is in your training or calibration data that should not be, and how did you check?
- Q115
Who is harmed if this system is deployed and is wrong in the way it is most likely to be wrong?
Mathematics and theoretical work
- Q116
Which step of the main proof is doing the real work, and where would an error most plausibly hide?
- Q117
What breaks if you drop [ASSUMPTION]? Is the result false, or only unproven?
- Q118
Is your result tight, and if not, where do you expect the true bound to be?
- Q119
What is the smallest non-trivial example, and does it behave as your theorem says?
- Q120
Which existing result does yours contain as a special case, and which does it fail to?
- Q121
What did you try that did not work, and what did the failure tell you?
- Q122
What would a computational check of this claim look like, and has anybody done one?
Humanities and text-based research
- Q123
Why these texts, these editions and this archive rather than others available to you?
- Q124
What is the counter-reading of your central passage, and why do you reject it?
- Q125
Where does your argument depend on a translation choice, and who made it?
- Q126
What is the status of the evidence you cannot access, and how does it constrain the claim?
- Q127
Which part of your argument is interpretation, and which is a claim about what happened?
- Q128
What is the strongest objection from the theoretical tradition you are not working in?
- Q129
What would a reader who does not share your framework take from chapter [N]?
Law and legal research
- Q130
Is your claim doctrinal, normative or empirical? Point to where the thesis changes register.
- Q131
Which authority is your argument most exposed on, and what happens if it is overruled?
- Q132
How does your reading survive the strongest contrary line of authority?
- Q133
What is the comparative work doing: illustration, or an argument that the transplant would function?
- Q134
Whose interests does your proposal cost, and have you stated the cost as clearly as the benefit?
- Q135
What is the institutional mechanism by which your recommendation would actually be adopted?
- Q136
Where does your normative claim rest on an empirical assumption you have not evidenced?
Education and practice-based social research
- Q137
What did you measure, and what did you want to change? Say where those two differ.
- Q138
How do you separate the effect of the intervention from the effect of the attention?
- Q139
What would have to be true for this to work in a setting with none of your advantages?
- Q140
What is the evidence that any change persisted beyond the end of the study?
- Q141
What is the practitioner supposed to do differently on Monday morning?
- Q142
What did the students or teachers in your study think you were doing, and does it matter?
- Q143
What is the cost of your intervention, in hours and in money, per person reached?
Practice-based arts and design research
- Q144
Where does the knowledge live: in the artefact, in the process, or in the written component?
- Q145
What can this practice show that an argument in prose could not?
- Q146
How would somebody else re-enact your process, and what would count as doing it correctly?
- Q147
What did the work refuse to do, and what did you learn from the refusal?
- Q148
Which decisions in the making were method, and which were taste? Defend the line you drew.
- Q149
How does the documentation stand in for the work, and what does it lose?
- Q150
What is the contribution for somebody who never sees the artefact?
Your R list
With JavaScript on, the questions you mark R collect here. In most theses that list runs to four to eight, and two of them are the same question wearing different clothes.
The log sheet
Five columns, and three of them matter more than the marks.
Keep this beside the bank on the first pass. Two pages, filled in as you go.
| Column | What goes in it |
|---|---|
| Q# | The question number |
| Mark | G · A · R |
| The gap, in your words | One line. What you cannot yet say |
| What fixes it | Rehearsal · Reread · Re-analysis · Ask supervisor |
| Fixed by | A date |
“The gap, in your words” stops an R becoming a vague dread, and “what fixes it” sorts the 20-minute problems from the three-day ones. “Fixed by” turns the list into a plan with dates, which is the difference between preparation and worry. The working copy ships it as a fillable table.
The three questions to check first
Three questions, worth more than the other ninety-seven.
If you have one hour rather than one afternoon, mark these three and nothing else.
- Q2 — The contribution, in one sentence, out loud.
- Q53 — The real limitation, not the written one.
- Q48 — The most plausible alternative explanation for your central result.
An amber or red on any of those three is worth more of your remaining time than the other 97 combined. They are the questions that get asked in every viva, in every discipline, and they are the three that cannot be answered by rereading a chapter.
What a mock does that a list cannot
It cannot hear your own voice.
This bank finds the questions you cannot answer.
It cannot tell you that your answer to Q47 was fluent, confident and did not answer the question, because you cannot hear that in your own voice.
Defense Readiness — a mock defence with a specialist in or near your field, and feedback within 48 hours of the session.
- Standard€599· one session
- Intensive€799· two rounds, feedback within 48 hours of each
- A second mock€225· a second committee member’s perspective on the same material
The Zero-Unknowns Guarantee
If you leave the mock with a question you cannot answer and we have not given you the route to answering it, the session is free. On Intensive, round two is free. Judged on the session itself, and set out in full at the guarantee.
The pre-defence bundle. Buy the Manuscript MRI alongside Defense Readiness and the MRI is €389 rather than its standalone €459, a reduction of €70. The Defense Readiness price is unchanged.
Or use the free route. A written strategy brief: tell us where you are stuck, and within 3 working days we send a three-step plan, including which service fits, or that none of ours does. No follow-up sequence.