49

For a PhD candidate that has done a fine job in their PhD, what are the common mistakes to avoid in a defense session?

  • 36
    Being too worried about it. I was pretty nervous, but in the end it was more of a stimulating discussion than anything else. – Steven Gubkin Sep 4 '15 at 21:33
  • 1
    As @StevenGubkin says, the point of the defense is for the committee to understand what you did. You're just giving a talk to a willing audience! – DanielSank Sep 5 '15 at 18:36
  • 2
    @Steven Not being nervous enough could also backfire. I've seen it happen. – Martin - マーチン Sep 8 '15 at 13:13
48

How the dissertation defense works varies enormously between fields, countries, universities, and departments. Any guidance has to take into account what the expectations are.

For example, in the defenses I'm familiar with, there are no common mistakes at all. In fact, I'm not sure I've ever seen anyone do anything worth calling a mistake. Here, the defense is basically a formality: you don't reach that point unless your advisor and readers are happy with your dissertation and confident you will easily pass, and in practice nobody fails unless something goes dreadfully wrong. That can happen (I know of one case in which a serious mistake was uncovered at the last minute, which was certainly embarrassing to both the candidate and the committee), but I've never actually been present when someone failed. So if someone from my department asked me about common mistakes they should avoid, I'd tell them not to worry about it: making it to the defense is the hard part, and then passing it is really straightforward. It's more of a celebratory ritual than an actual test.

But of course this depends on how the defense is handled in your department.

  • 1
    Where I did my PhD it was even further to the side you describe. Before the defense one gets an email with a formal letter attached stating that the thesis has been approved (i.e. stating that the work is worthy of a PhD). Essentially the only way to fail at the defense is if it turns out you did not do the work yourself. – Tobias Kildetoft Sep 5 '15 at 10:14
  • 4
    @Calchas - In case you can afford the time, I guess people in this thread would be willing to hear about the reasons why that bloke got a flat out "No". What exactly did he do to deserve it? – 299792458 Sep 6 '15 at 15:50
  • 1
    @Calchas - Haha, no but what I meant was, if the writing was already on the wall, how did things get to the stage of dissertation defence? – 299792458 Sep 7 '15 at 5:50
  • 2
    @TheDarkSide: maybe the university procedures don't really account for someone who wants to defend the indefensible, so if the candidate ignores all advice and insists, then they're entitled to submit and defend, and the only way to stop them would be to expel them first (which is even more hassle)? – Steve Jessop Sep 7 '15 at 11:24
  • 1
    @Mehrdad: I don't know the details (it was some years ago, and I wasn't directly involved). As I remember it, during the week before the defense one of the readers found a serious mistake in a central part of the work, which everything else depended on. It looked like it could plausibly be fixed, but it wasn't 100% clear whether or how it could. The student was reluctant to cancel the defense, and he worked like crazy to try to fix the mistake before the defense, but he didn't succeed and didn't pass. However, he passed on his second try, after spending a few months fixing everything. – Anonymous Mathematician Sep 7 '15 at 16:10
24

Let me turn the question around and answer what I want a PhD candidate to demonstrate during a defense. There are in reality only 3 things:

  • Demonstrate mastery of the subject, i.e., understand where the problem you researched comes from, why it is relevant, what others have already done, and why their approaches are insufficient.

  • That you have made a creative contribution to resolving the question you worked on. The creative contribution may really have been only a single idea or insight, grounded in an understanding of the field. In reality, you will have likely developed it jointly with your adviser.

  • That you have followed up this idea with diligent, comprehensive work that explored the potential and consequences of the idea.

If you cover these three topics, you will be on the good side.

Now for the common mistakes: Among the bad or difficult PhD defenses I have sat through -- of which unfortunately there were a number -- the candidate typically covered only one or two of the areas above. I can think of defenses where a candidate demonstrated good, diligent work, but was almost entirely unaware of the scientific background of the work. I can also think of defenses where the candidate understood the field well, but all of the work was essentially applying existing methods to problems to which it had previously been applied. And I can think of defenses where the candidate had had an idea, but whether the idea was worthwhile pursuing or not had not been explored in any kind of detail through follow-up work. In other words, I've seen PhD candidates make all kinds of mistakes. In reality, I think that the problem was in all of these cases not with the defense, but with the candidate, as usually becomes apparent during the Q&A session at the end.

  • 1
    Just for my own sake so I don't screw up when it's my turn -- how do you want the balance to be between the presentation and the dissertation? For example, if a dissertation covers all 3 but the presentation is only supposed to be 30 minutes so it focuses on 1 or 2 of the areas, is that a mistake? It seems hard to cover any one of the areas in a short presentation while all 3 seem possible in the full text. – tpg2114 Sep 5 '15 at 4:37
  • 11
    The mistakes you list also seem to indicate that the candidate's advisor has done a poor job. – Marc Claesen Sep 5 '15 at 5:23
  • @tpg2114 I guess it depends on the form the defense takes, but where I am i would see it as the examiners' job not to finish the defence until they are sufficiently satisfied with all 3. I guess you could mess up by de-railing all their questions, but I'd expect that to indicate one of the 3 was missing anyway. – Jessica B Sep 5 '15 at 6:01
  • 1
    @tpg2114: Have a brief introduction (maybe 5 minutes) of the field and the question you're looking at, and then spend the rest of your time on the other two points. Your mastery of the subject will become apparent throughout the defense and the questions. – Wolfgang Bangerth Sep 5 '15 at 12:57
  • @MarcClaesen: Yes. As well as cases where a student forced the issue because they had a job lined up and requested the defense even though the adviser was not convinced it was time yet. – Wolfgang Bangerth Sep 5 '15 at 12:58
19

IMO, the most common is being defensive (perhaps encouraged by the name of the ritual). For instance if a member asks "Why did you use Smith's procedure...", and the candidate starts backtracking and apologizing, saying "I suppose I could have used Jones' procedure, that would have been better", when the question simply asks you to explain why you used Smith's procedure.

9

In the end, the person in the room who probably knows most about the work done is the candidate.

  • Don't get nervous
  • Make sure your slides cover what you did, in sufficient detail. Don't overdo it, your time is limited. Make sure (e.g. by rehearsing) you fit in the allotted time. Might have to only show a slice for this, select it carefully.
  • When asked about stuff outside your direct work, it can be better to say you don't know than to try to come up with an explanation on the spot.
  • Check possible questions (within reason!) and have answers ready
  • 2
    All due respect, but bullet 4 is easier said than done. You can't anticipate everything. – 299792458 Sep 6 '15 at 15:42
6

I have been to multiple defenses, some common/general mistakes:

  • Not knowing why they use certain methods as opposed to others. Meaning that you can answer what you are doing and how the formula works, but you can't answer why not use X method instead.

  • Not knowing the specifics of certain methods/formulas and their reasons. Ex: we take X from Y, and you can't answer why you can/should take X from Z. more realistically: we take a measurement from the center, why not the top or bottom?

  • Charts with too little information, and trying to make connections that "are there" in most datasets, but doesn't imply they will be in yours. Don't create curved lines if you are missing dots.

  • Not knowing what is the importance of whatever it is that you are doing. Can you answer the question: "How is this useful for someone else?"

  • Answers in the following form: why are you using "matlab?", you: because my advisor said so. Expected response: because my advisor has plenty of experience with it and we know it can solve the problem better than other software.

1

The worst mistake you can make is to insist on defending on a certain date, over your advisor's objections. (I have seen this happen.)


If I may be forgiven a small anecdote: My aunt was asked a question in her defense (physics) that she didn't know how to answer. She said to the questioner, "Perhaps a sketch will help make this clear," and attempted to draw some coordinate axes... which didn't come out straight... and after several erasures and retries, the questioner got fed up and said, "It's okay, I know you know it, let's move on."

  • 4
    I'm sorry, but I didn't follow the point of your anecdote. Could you please make that explicit? It seems to give the impression that you are suggesting derailment. – 299792458 Sep 6 '15 at 15:45
  • 1
    @TheDarkSide - I don't understand what or who might be getting derailed... but the anecdote is supposed to give an example of a committee wanting to help the baby bird get out of the nest (i.e. the department). – aparente001 Sep 6 '15 at 18:03
1

A common mistake to avoid is getting too nervous or defensive, and trying to cover up the limits of your knowledge. The committee will try to explore the limits of your knowledge and keep pressing for answers to increasingly difficult questions until you say "I don't know." Be expecting this. It's OK to not know the answer, especially when the question is not core to your thesis. If you can identify why a question is not core to your thesis, and articulate that clearly, you get bonus points for that (as long as you do it tactfully and don't imply that it's a bad question). Also, depending on the committee, they may be looking for a better answer like "I don't know, but for X reason(s) I suspect the answer might be Y and I would go about finding the answer by Z." If the question is core to your thesis, offer to find out the answer and get back to them (especially if they're on your committee).

The better your presentation, the more questions you're likely to get that try to expand outward into territory you haven't prepared for. Take that as a good sign!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.