Questions to Ask at a Dissertation Defense
For committee members, external examiners and people in the audience looking for questions to ask at a dissertation defense, and for candidates who want to know what is coming. The list follows the order a defense tends to take: opening questions, then the research question and its framing, the method, the findings, the limits, and the contribution and what comes after it. Formats differ between universities and countries, from a public doctoral defense to a closed viva or a master's thesis defense, so check with the chair how questions are taken where you are.
Want questions from the whole vault instead? Try the random question generator.
The questions
Each question, and why to ask it
Opening
Can you summarize your dissertation in a few minutes for someone outside your field?
Why ask it
Most defenses open with some version of this, and it tells the room whether the candidate can see the whole project or only its chapters. A strong answer gives the question, the approach and the main finding in that order, without the chapter-by-chapter tour. Candidates: rehearse it aloud until it fits in three minutes.
What drew you to this topic in the first place?
Why ask it
A warm-up that settles nerves, and the answer often explains choices that look odd later, such as why one site or one population was chosen. If the story is purely personal, follow with what the field stood to gain from the question being answered.
If we remember one finding from this work, which should it be?
Why ask it
Forcing a ranking shows whether the candidate knows which result carries the dissertation. Note the answer down: it is the claim that the method and limits questions later in the session should test hardest.
What surprised you most during the research?
Why ask it
Someone who did the work has a ready answer, usually a result that ran against expectation or a source that would not cooperate. 'Nothing, really' is worth a gentle second try, because a project with no surprises may have found only what it set out to find.
How has your thinking changed since the proposal?
Why ask it
Committee members who sat through the proposal can compare the answer with their own memory of it. Changes are normal and usually a good sign. What you are listening for is whether each one was a decision with a reason or a drift nobody recorded.
Which part of the dissertation are you proudest of?
Why ask it
It hands the candidate firm ground early, which makes the harder questions go better for everyone. The choice is informative too: some name a result, some a method they built, some a chapter that took four rewrites.
What was the hardest part of the project, and how did you get through it?
Why ask it
Friends, family and newer students in the audience can ask this without having read a page. Replies range from a middle section rewritten from scratch to a year when recruitment stalled. Examiners can use it too: if the trouble sat in one chapter, ask later what it left unresolved there.
Framing
What is your central research question, and how did it come to be worded that way?
Why ask it
Ask for it in a single sentence before the history. If the spoken version differs from the one in chapter one, find out which of the two the study actually answers, since the conclusion can only be judged against one of them.
What gap in the literature does this study fill?
Why ask it
Nearly every candidate is asked this, so expect a prepared reply. The useful follow-up is why the gap existed: nobody thought of it, nobody could get the data, or people judged it not worth doing. Each reason sets a different bar for the contribution.
Why did you choose this theoretical framework over the alternatives?
Why ask it
A good answer names at least one rival framework and says what it would have hidden or shown instead. A framework inherited from the advisor's own work is fine, but the candidate should still be able to argue for it in their own words.
Which authors or studies shaped your thinking most, and where do you part company with them?
Why ask it
The second half is the real question. A candidate who agrees with every source has summarized a literature. One who can say where a major author is wrong or incomplete has taken a position in it.
How do you define your key concept, and what changes if a reader defines it differently?
Why ask it
Swap in the actual term when you ask, whether that is resilience, engagement or risk. Many disagreements at a defense turn out to be two people using one word in two ways, and settling the definition early saves time in the method and findings rounds.
Who would disagree with how you have framed the problem, and what would they say?
Why ask it
This invites the candidate to make the opposing case fairly, which is harder than rebutting it. If no critic comes to mind, an examiner from a neighboring field can describe how their own discipline would have put the problem and ask for a response.
What did you decide to leave outside the scope of the study, and why?
Why ask it
Scope decisions are where a project becomes finishable, so treat the answer as a record of judgment and not a confession. Press only where something left out looks able to change the main finding.
What has been published since you finished your literature review that you would now include?
Why ask it
Literature chapters are often written years before the defense. One or two recent titles and a sentence on how each bears on the findings is plenty. Examiners who have a particular paper in mind should name it and not make the candidate guess.
What assumptions does the study rest on, and which one worries you most?
Why ask it
Every design takes something as given: that people report their own behavior accurately, that an archive kept what mattered, that a model's conditions hold in the field. The one the candidate calls shakiest is a preview of the limitations discussion, so hold the answer until then and see whether the written chapter owns up to it too.
Method
Why is this design the right one for your question, and what did you consider and reject?
Why ask it
The rejected options are the test. 'It is what my lab does' is honest but incomplete, so ask what a different design would have cost and what it might have shown.
How did you select your sample, sources or cases, and who or what is missing from them?
Why ask it
Adapt the wording to the field: participants, archives, texts, sites, specimens. Whatever is missing sets the edge of every claim in the conclusion, so come back to this answer when the discussion reaches generalization.
How did you decide you had enough data: enough participants, cases, sources or runs?
Why ask it
The reply depends on the tradition: a power calculation, saturation in interviews, every surviving document in a collection, or the months the funding allowed. Any of those can be defended. The one to follow up is a number set by convenience and described afterwards as a principle.
How do you know your measures capture what you say they capture?
Why ask it
In a survey study this is about validity and reliability evidence. In interview or archival work it becomes how the candidate knows a source means what they take it to mean. A reply that rests only on 'earlier studies used the same scale' deserves one more question about whether it suited this population.
Can you walk us through one piece of the analysis, from the raw data to the claim in the text?
Why ask it
Pick a specific claim and give its page number. One thread followed from transcript, dataset or document to the sentence in the chapter says more about rigor than a general account of the method. It is also where it becomes plain which steps the candidate did by hand and which were left to software or a collaborator.
Which decision in the analysis could reasonably have gone another way, and did you check the other way?
Why ask it
Coding schemes, exclusion rules, model specifications and cut-off points all involve choices. You are asking whether the main result survives a different reasonable choice. If that was never checked, it may be a revision request and not a failure, depending on how much rests on it.
What went wrong during data collection, and how did you adapt?
Why ask it
Something always does: low recruitment, a closed archive, broken equipment, a site that withdrew. The adaptation matters more than the mishap. Check that the change is written into the methods chapter and not only told in the room.
How did your own position or assumptions shape what you collected and how you read it?
Why ask it
Standard in qualitative and community-based work, and worth asking elsewhere too, where it becomes a question about which results the researcher was hoping for. Look for a concrete instance, such as an interview that went differently because of who was asking.
What ethical issues came up, and how did you handle them?
Why ask it
Formal approval is the starting point, and the rules for it depend on the institution and the country, so ask what applied there. The more telling part is what arose that the application did not foresee: a participant who wanted to withdraw, data that identified someone, a finding that could harm the people studied.
Could another researcher repeat this study from what is written in the methods chapter?
Why ask it
If the honest answer is 'not quite', ask what is missing: the interview guide, the code, the search terms, the exclusion criteria. Those can usually go in an appendix, which makes this one of the easier revisions to request and to complete.
How did you handle data that did not fit: missing responses, outliers, or cases that contradicted the pattern?
Why ask it
Every dataset has some, so 'there were none' is the reply to probe. Ask how many cases were dropped or set aside and whether the main result holds with them left in. In qualitative work, ask for the participant or document that argued against the main theme.
Findings
Which result is the strongest, and what is it resting on?
Why ask it
Have the candidate point to the exact table, passage or figure. Then compare it with the finding they asked the room to remember at the start. When the best-supported result and the headline result turn out to be different ones, ask which of them the abstract should lead with.
Which result are you least sure of?
Why ask it
Most candidates know exactly which one, and saying so aloud is a mark of maturity, not weakness. The follow-up is how much of the conclusion leans on it and whether the text signals that uncertainty to a reader.
Can you take us through this figure or table and tell us what we should see in it?
Why ask it
Name the figure and give everyone a moment to find it. If the display is dense, suggest starting with the axes or the column headings, since a figure made many months ago can take its own author a second look. It also gives audience members who have not read the document a way into the discussion.
What else could explain your main result, and how did you rule it out?
Why ask it
Asked in some form in most fields, so candidates do well to arrive with two or three rivals already listed. The best replies bring up an explanation the examiners had not thought of. Where one cannot be ruled out with the data in hand, the fair outcome is that it gets named in the limitations.
How do your findings sit against earlier work: where do they confirm it, and where do they contradict it?
Why ask it
Ask for one of each. When a result contradicts an established study, the candidate should be able to offer a reason, such as a different population, period or measure, and not simply report that the two disagree.
Is there a result you expected and did not find?
Why ask it
Null and absent results tend to get a paragraph when they deserve a section. Hearing whether the candidate treats one as a finding or as an embarrassment shows how they will handle it in print.
Is the effect large enough to matter in practice, or only large enough to detect?
Why ask it
For quantitative work, have it put in everyday units: points on a test, days in hospital, dollars per household. In studies without numbers the parallel question is how common or how central a theme was across the material.
What is the thread that ties the chapters into one argument?
Why ask it
Especially useful for a dissertation built from separate papers, where each chapter was written to stand alone. The candidate should be able to say what the whole shows that no single chapter does. If they cannot, the introduction or the conclusion probably needs another pass.
Limits
What do you see as the biggest limitation of this study?
Why ask it
Candidates should arrive with a real one and not 'the sample could have been larger'. The answer to listen for names the limitation, says which conclusion it weakens, and says why the work can still be trusted.
How far can these findings be generalized, and where would you not apply them?
Why ask it
Tie this back to who or what was missing from the sample. The reply to hope for draws the boundary unprompted: one named setting, period or group the results should not be stretched to. Then check that the conclusion chapter is as careful as the spoken answer.
If you started again tomorrow, what would you do differently?
Why ask it
Hindsight is the point of this one. A specific change, such as piloting the instrument or starting recruitment a term earlier, is worth more than a general wish for more time. 'Nothing' usually means the candidate heard a trap, so say plainly that it is not one and ask again.
What evidence would make you give up your main conclusion?
Why ask it
A conclusion that nothing could overturn is not being held on the evidence. Precise replies sound like 'a failed replication at a second site' or 'a letter showing the opposite'. A vague one is the cue to ask which single piece of evidence the claim leans on most.
Which claim in the conclusion goes furthest beyond your data?
Why ask it
Final chapters are often written last and fastest, and they are where recommendations outrun results. Have a sentence from the conclusion ready to quote. The usual fix is a change of verb, from 'shows' to 'suggests', and it is a small one to agree on.
Which limitations came from the time and money you had, and which from the design itself?
Why ask it
The first kind is forgivable in student research and the second kind is the one to discuss. The split also separates what a funded follow-up could fix from what would need a different study altogether.
Contribution and next steps
What is the original contribution of this work, in one or two sentences?
Why ask it
For a doctorate this is usually the bar the work is judged against, and a master's thesis is often held to a different one, so examiners new to the institution should read its regulations first. The answer should claim something specific: a new dataset, a method, a corrected finding, an argument nobody had made.
Who should do something differently because of what you found?
Why ask it
Teachers, clinicians, engineers, policymakers, archivists: make the candidate pick. Then ask what exactly that person should change on Monday, which exposes whether the implication is real or decorative. For purely theoretical work, ask which scholars should now change a line of argument.
What does this work add to the theory itself, beyond one more case of it applying?
Why ask it
Many dissertations apply an existing framework to a new setting and stop there. Asking what the setting taught the framework, where it bent or failed, pushes toward the stronger claim and often produces the best paragraph of the final revision.
What question does your dissertation raise that it cannot answer?
Why ask it
Good research ends with a better question than it began with. This one suits audience members well, since it needs no knowledge of the method and candidates usually enjoy answering it.
What is the next study that should follow from this one?
Why ask it
Listen for a design and not a topic: who would be studied, how, and what result would be interesting. A candidate heading for a postdoc or a faculty post will need this answer in interviews, so the defense is a useful first run.
Which parts do you plan to publish, and where?
Why ask it
Advisors and committee members can turn the reply into practical help by suggesting a journal, a press or a co-author. If the candidate is leaving research, ask what should happen to the data and findings so they are not lost.
What do you know now about doing research that you did not know when you started?
Why ask it
A closing question that lets the candidate speak as a researcher and not as a student. Answers tend to be about patience, scoping or asking for help sooner, and the newer students in the audience are often the ones who gain most from hearing them.
Is there anything we have not asked that you hoped we would?
Why ask it
Chairs often end with this. It gives the candidate one chance to show work that nobody reached, so candidates should have something ready: a favorite finding, a detail of the method, a plan. 'No, thank you' is also a perfectly good answer at the end of a long session.
Asking and answering at a defense
Practical guidance for the conversation itself
Before the day
Read with page numbers in hand
Questions tied to a page land better than questions from memory. As you read, mark the sentence each question comes from and bring the list with the references on it. Five or six prepared questions is usually enough for one examiner, because the candidate's answers will supply the rest.
Find out how the room is run
Defenses differ widely. Some are public events with an audience, some are closed examinations, and the order of questioners, the time each gets and whether visitors may ask anything are set locally. The chair or the graduate office can tell you how it works at that university, and it is better asked a week ahead than at the door.
Know what your seat is for
Roles differ from one system to the next, so ask the chair what yours is. Where the advisor sits on the panel, they have had years to ask and usually leave most of the questioning to others, and in some countries the supervisor does not examine at all. Committee members test the parts nearest their own expertise, an external examiner holds the work to the standard of the wider field, and an audience member is a guest who asks only if the chair opens the floor. Choose from the list accordingly: Method and Findings for specialists, Opening and the closing group for everyone else.
If you are the one defending
Answer the Opening and Limits groups aloud, to another person, at least once. Those are the questions that arrive almost regardless of topic, and spoken answers come out differently from the ones in your head. Then ask your advisor which chapter each committee member is likely to go after, and reread those pages last.
Putting the question
One question, then stop
A question with three parts gets an answer to the last part. Ask the one thing, let it be answered, and use a follow-up for the next. The candidate is already holding the whole document in their head.
Say why you are asking
'I ask because the sample in chapter three is all from one district' gives the candidate something to aim at. Without the reason, a nervous person will often guess at a hidden objection and answer that instead.
Ask from outside your field without apology
Committee members from other departments and visitors in the audience often ask the most useful questions, because those have to be answered without jargon. 'Explain to me why that matters' is a legitimate examination question.
Let the silence run
A pause of several seconds while a candidate thinks feels long from the examiner's chair and is entirely normal. Rephrasing too soon turns one question into two. If the pause goes on, offer to come back to it after the next examiner.
Judging what you hear
A good answer returns to the evidence
Listen for the candidate pointing back into the document: a table, a quotation, a chapter. Answers that stay at the level of general principle may be right, but they do not show command of this particular study.
Conceding a point is a strength
A candidate who says 'you are right, the claim on that page is too strong' is doing what researchers do. One who defends every sentence equally hard makes it difficult to tell which claims they would stake the work on.
'I do not know' can be a complete answer
What follows it is the test: how the candidate would find out, or why this study could not address it. For candidates, that is nearly always better received than a guess delivered with confidence.
Separate the person from the document
Nerves make fluent people halting. If a spoken answer is weak but the written chapter handles the point well, the chapter is what is being examined. Say so when the committee deliberates.
What sours a defense
The lecture with a question mark
Two minutes on your own work followed by 'would you agree?' examines nobody. If you cannot state the question in two sentences, it is probably a comment, and comments can go in a written report, where the examiners file one, or wait until the room has emptied.
Examining the thesis you would have written
Every examiner would have designed the study differently. The fair test is whether the chosen design answers the stated question, not whether it matches your preference. Raise the alternative once and move on.
Springing a fatal flaw on the day
A problem serious enough to threaten the result should reach the advisor or the chair before the defense wherever the local rules allow that, so ask what the procedure is. Saving it for the room helps no one, least of all a candidate whose family is sitting in the third row.
Arguing with a colleague through the candidate
Old disagreements between committee members sometimes surface as questions. If you notice you are watching for another examiner's reaction and not the candidate's answer, let the question go.