Outline for presentations at final exams ("defenses") of doctoral dissertations
Neil C. Rowe, 6/09
(a) Defenses usually consist of a presentation no longer than 45 minutes, followed by questioning from one to two hours. Expect to be interrupted for short clarification questions during the presentation. Questions after the presentation typically explore the assumptions, limitations, extensions, and applications of the dissertation work.
(b) The defense is intended to be a "public" presentation. That means you should design it not for your doctoral committee but for intelligent listeners in the field in which you are getting the Ph.D. Avoid acronyms and other jargon as much as possible. Your committee's vote on your performance will put significant weight on how you handled questions from nonmembers of the committee.
(c) As with presentations at science and engineering conferences, you should rehearse your talk in advance. It is best to give a practice run for your entire dissertation committee, but at least you should present it orally to your dissertation supervisor. The practice run should help both with judging the length of the talk and finding things that can be improved.
(e) The core of your presentation should be a set of novel claims from your work and the validation of your claims. This is supplemented with how these claims relate to prior work and what is different about them, plus speculations about future implications of what you have done.
NOTE: Slide counts recommended below are for 28 point font. Do not use font smaller than this for exams with videoconferencing, and less than 20 point otherwise, with the possible exception of important figures and tables that cannot be compressed.
1. What problem are you addressing (1-2 slides). Focus on the primary problem if there is more than one.
2. Why this problem is important (1 slide).
3. What contributions you have made that no one previously has done (1-2 slides). The contributions must be to the field of the degree. State them as claims. Most of the remainder of your presentation will be the validation of your claims.
4. Previous work addressing the same problem with different methods than yours (1-3 slides, depending on the topic). Give names of researchers and summarize succinctly what they did. Explain why previous work didn't solve your problem completely.
5. Previous work addressing different problems with similar techniques to those you used in the dissertation work (2-5 slides, depending on the topic). This can be short if you used well-known techniques.
6. Design of the validation for each of your claims (5-15 slides). The validation can include experiments, tests, and proofs. If you built something, this is where you describe it. Note that validation can be of design as well as of implementation, although validation of an implementation is more convincing. Thorough validation is the key feature distinguishing Ph.D. dissertations from Master's theses.
7. Results of the validation of each of your claims (2-10 slides). Usually this is presented in the form of statistics and some analysis of data that has been collected.
8. Conclusions: How your validated claims have contributed to the solution of the original problem (1-2 slides).
9. Broader implications: Why your work is useful to society (1-2 slides).
Total: 15-42 slides