Guidelines for defending your Ph.D. proposal or dissertation.
So, you’re finally at the point when you are about to defend your Ph.D. dissertation. In the United States, this is generally a two-stage process: first you write and orally defend your dissertation proposal (the proposal exam/defense), and then you turn your proposal into a complete dissertation and defend that in an oral exam (the final or dissertation exam/defense). These stages are typically at least a year apart, and sometimes longer. In other places, notably Sweden, there is no proposal and you skip ahead to the Ph.D. dissertation writing and defense straight away. For all of these situations, I have some general guidance I give all my students — read on to find out what.
Disclaimer: This a personal blog, so these are obviously my own personal opinions. Feel free to agree or disagree with any of this!
The Dissertation Proposal
The U.S. dissertation proposal has some special considerations that are worth discussing on their own. In general, the idea of a dissertation proposal is a fantastic one: it establishes what essentially amounts to a contract between the Ph.D. student and their advisor and committee. The idea is that if the student fulfills the contract, the committee will award them a Ph.D. Thinking about the proposal in this manner means that the preliminary exam, in which the student defends the proposal, can be thought of as a contract negotiation. In other words, during the preliminary exam, the two parties engage in a lively scientific discussion of what the contents of the Ph.D. dissertation being written should be.
Of course, if the proposal is a contract, changing it means changing the contract. While this is typically fine — everyone knows that predicting the future is hard and that science is risky business — you will want to inform your committee whenever this happens so that they are not caught by surprise when they see your final dissertation. In other words, if you change your plans, you shouldn’t have to hold another proposal defense, but you should definitely reach out to your committee members — and potentially meet with them separately or as a group — if the changes are substantial.
Here’s where the contract metaphor breaks down: working on the contract rather than the work itself is counterproductive, but remember that the proposal document is the beginning of your Ph.D. dissertation. This is another great advantage of the proposal system: Any time spent on this task will help the future you in completing your dissertation. Do yourself a favor and use the time to write a good proposal document that is the seed of a great dissertation.
One final note about the proposal document: you should clearly differentiate between the work you have already done and the new work that you are proposing to do. This is also the only difference between the dissertation proposal and the dissertation: they may both have the same chapters, but in the former, some of these chapters are unfinished and lack results, whereas in the latter, all chapters are complete (with the exception of the future work chapter, which can be thought of as research questions with no answers). Because this is the only difference between the two documents, all of the remaining guidance in this document applies equally to the proposal as well as the final dissertation.
Credit Where Credit Is Due
In many disciplines, including computer and information science, it is common to have many authors on research papers: not just you and your advisor, but sometimes additional students, postdocs, and collaborators. Practically speaking, this means that several students may include a particular paper in their dissertation. For most U.S. universities this is fine and in fact the norm, but you still want to tread carefully: obviously, every student should be getting their Ph.D. degrees based on their own efforts, and not by getting a free ride from others.
To eliminate any questions about who did what for a multi-author paper, it is often a good idea to include a section in your introduction chapter where you enumerate all of the papers you include in your dissertation and which includes a statement on your role and contributions to each one. A rule of thumb is that most, if not all, of the papers you include in your dissertation are ones where you had a substantial intellectual contribution to the work. In other words, you should not just have helped out with the user study sessions or wrote a few lines of code. Ideally you were part of formulating the research idea and developing the research from the beginning.
Disseminating the Document
Now for some very practical advice on disseminating your proposal or dissertation proposal to your committee members: the rule of thumb is to do it early, but allow for revisions. In most U.S. universities, there is a minimum period of time — typically on the order of ten business days or three weeks — for a student to provide a document to their committee in advance of the exam. In a perfect world your document will be completely ready in time for this deadline. In the actual world we live in, this is rarely the case. Still, I find it is better to be safe and adhere to the deadline by sending a draft in time for the deadline but to continue revising the document up to perhaps a week before the exam. Be sure to communicate with your committee if this is the case so that they are aware that the document may be changing, but don’t worry about it too much. Many faculty members will not look at a document — save for an initial once-over when they receive your email — until perhaps a week before the exam anyway (maybe I am telling on myself here?).
Another practical tip for when you are still revising the document inside the period leading up to your exam: don’t send your document as a PDF attachment in an email, but instead send a link to your website, Dropbox, or Google Drive. This will allow you to quietly update the PDF in the intervening time from your initial dissemination email to the exam without having to keep sending PDF attachments to your committee. Also, it prevents you from cluttering up your committee’s inboxes with huge email attachments! In fact, one of my former faculty colleagues has a script that automatically discards emails that have a large attachment.
Get your exam time on the books as soon as possible. This may take a while to schedule, so be sure to allow for some time to get this done and don’t put it off for too long. Given the above constraints on how much time your committee needs to review a document, you should probably start the scheduling process a full semester in advance. In other words, if you want to schedule your exam at the end of the spring semester, you should likely start the process at the beginning of the spring semester. If you are going for the beginning of the spring semester, you probably will want to start the conversation halfway through the fall semester.
Remember that your department most likely needs to know about the time and place of your exam for administrative purposes. In other words, don’t just settle on a time with your committee, but communicate with the Ph.D. coordinator or director at your department to ensure that you are meeting all of the formal requirements. Some universities still require approval for involving remote committee members in a Ph.D. dissertation exam, for example. In practice, all of these administrative details are your responsibility even if they formally may belong to your advisor — remember who will ultimately lose if they don’t get done.
Time zones and video conferencing links can be confusing to find in the heat of a busy day for your committee members, so I recommend sending actual calendar invitations with all of the details to your committee members.
The night before the defense, you should send out a reminder to your committee about your impending defense. Professors are busy people, and, again, the person who suffers the most from one of your committee members forgetting the exam is you, not the committee member. You obviously don’t want to inconvenience your committee members by sending incessant reminders about your exam, but a reminder at the beginning of the week and the night before is certainly within reason. Include a link to your document (a link, not the PDF itself, remember) and the time and place (and videoconferencing information).
Giving the Talk
All right, the big day is finally here, and you are about to actually go ahead and do the thing: defend your proposal or dissertation. Great! The first rule is to relax. Exams are high-stress situations, but, trust me, you would not be at this point if you were not ready (advisors have a way of protecting you this way). The best way to think about the exam is not really as an exam or a defense — these terms are really too confrontational — but rather as a conversation between experts. Yes, when you started out as a fresh Ph.D. student you were certainly a greenhorn with a lot to learn, but look at how far you have come! Try to enjoy the experience, stressful as it can be.
All proposal and Ph.D. dissertation defenses begin with a talk where the candidate presents the work in the document. My rule of thumb, which I realize can be controversial and not shared among all advisors, is that your defense talk should not be longer than 30 minutes. There is nothing worse in a Ph.D. defense than a candidate who just drones on and on and squanders all of the limited time available for questions and discussion. Many committee members will schedule only an hour for a defense, and few will allow for more than two hours. Their time is valuable. Also, remember that they are all supposed to have read your document! In other words, there is no need to belabor every single point in your dissertation during your talk. Instead, think of your defense talk as an overview that focuses on the key findings, technical innovations, and general takeaways. Don’t waste too much of your limited time talking about unimportant details, but zero in on the novel ideas and findings in your thesis.
For a proposal defense, what you are really looking for is feedback on the upcoming work, so make sure that you leave ample time for detailing your plans and listing the questions you need help with. You probably should include a timeline for the remainder of your Ph.D. studies as well.
Obviously, after a couple of years as a Ph.D. student, you have a lot more material than 30 minutes and it can be tempting to want to show off all of this knowledge. Resist this temptation! Long-winded talks during which committee members squirm and eye their watches can be agonizing. Instead, move your additional slides to the end of your presentation after the closing slide. Answering a committee member’s question with a backup slide that you prepared in advance is a power move that is sure to impress.
Traditions vary between universities, but generally speaking you should be expecting questions during your talk and not just at the end — sometimes many of them. This is another reason why aiming for 30 minutes or less is a good idea: counting the questions and answers, your talk will likely go to 45 minutes or more.
Try not to get disheartened if you are getting a barrage of questions: this is not an indication that you in any way are failing the exam. Instead you should see it as an indication that your audience is interested in your work and is engaging with it to learn more.
Remember that you are here because you are an expert in this topic, and use this as a guidance for how to answer each question. An expert takes their time in answering questions. An expert asks for clarification if something is unclear. An expert answers questions in a precise and concise manner, checks with the questioner if the answer is satisfactory, and develops it further as necessary. An expert does not panic, does not freeze up, and does not go on long, rambling tirades with only tangential relation to the question. If you find yourself rambling, cut yourself short and check with the questioner. Finally, don’t be afraid to say that you don’t know the answer to a question, or that a question is outside the scope of your work. Only fools think they know everything; experts know precisely what they don’t know.
As already stated, the defense is a stressful event, and it can be hard enough just answering questions — how can you ever be expected to also remember them afterwards? Obviously you cannot, so you should be taking notes of what is being said and the feedback you are receiving. This will come in handy when you are revising the document in the days and weeks to come. Of course, taking notes in the heat of combat can be difficult as well, so don’t hesitate to ask your advisor — who is supposed to be on your side, or at least a somewhat neutral party — to do this for you. Ask them to summarize each of the questions as well as the feedback and comments received. In my experience, this kind of document is also a great way for your advisor to summarize the required revisions requested by your committee. Alternatively, you could also record the defense for later review, but this could be problematic since the closed sessions may be regarded as private. Be sure to check with your committee if you want to do this.
What Happens Next?
Most U.S. proposal and dissertation exams are organized into four stages: the talk, the public Q&A with the audience, the closed Q&A with only the student and the committee, and the closed committee deliberations. Once you have answered all of the committee’s questions during the closed Q&A, they will send you out to stew alone while they talk about the written dissertation and your oral performance. Closed committee deliberations are usually a conversation between committee members where they take turns discussing the strengths and weaknesses of your case.
Try not to read too much into how much time the committee is taking for their discussion. They may be engaged in deep and troubling discussions about your work and how to fix it, sure, but it is more likely that they are just chatting about the weather and catching up. Maybe they are discussing how to best nominate your work for an award! You have no way of knowing, so don’t worry about it.
In fact, for U.S. universities, the most common outcome for both a proposal or a Ph.D. defense is revisions required anyway, so it is better to make your peace with this in advance. Yes, it is certainly possible — and I have seen it happen a few times — that your document is near-perfect and just requires a big stamp of approval from the committee. It is more likely, however, that there are any number of big and small revisions to be made, and you will be spending the next few days, weeks, or months making them. That’s fine: your document will be better for it.
The real question is the form of these revisions. Realistically, the best outcome is that the revisions are sufficiently minor that only your advisor is required to check them. In this situation, the rest of the committee will typically sign off on the exam and the dissertation, with only your advisor holding off with their signature until you have made the revisions to their satisfaction. A worse, but certainly not catastrophic, outcome is when several committee members want to hold off pending the opportunity to review your changes. The worst case is probably where the revisions are so significant that a new exam is necessary, but this is a rare occurrence in my experience.
In general, the U.S. system with revisions being the norm rather than the exception is a good thing for science because it leads to higher quality dissertations. In contrast, for my Swedish Ph.D. defense, I was required to submit my dissertation for printing several weeks prior to the defense date, which in practice meant that none of what was discussed during the Ph.D. defense actually made it into the document. However, the bad part of the U.S. system is that the defense can easily become anticlimactic. You did all this work, defended it in front of your peers, and emerged unscathed at the other side only to… spend the next three weeks fixing typos? Hardly the glorious end to a Ph.D. journey of half a decade that you had anticipated, right?
Regardless, at this point in the process, the ball is in your court. I know that it can be painful to have to go back to the word processor once again, but my recommendation is to get on with the revisions as soon as you can and while the feedback is fresh in your mind. Hopefully you can get all the changes done and approved, the paperwork signed and submitted, and the Ph.D. awarded in no time at all! Good luck!