Question One:
How do I find a thesis advisor or topic?
Answer:
The initiative for finding a thesis advisor lies with the student. The student should talk to all individuals whom he or she thinks might be potential thesis advisors. It is not necessary to know the person or to have had the person as an instructor in class, although of course, discussion will probably begin more smoothly with those individuals you know. The topic for a thesis can either be a problem formulated by the student or may be a problem of interest to the advisor. It is perfectly appropriate to talk to individuals and ask whether they are interested in thesis advising and, if so, what topics they might have an interest in. Everyone in the Graduate Program in Biostatistics is on the graduate faculty. Since the relationship between thesis advisor and student is necessarily a very close and somewhat stressful relationship, there is no obligation upon either the student or the professor to join forces. It is not part of the faculty member's duty to advise theses for a particular person. Conversely, a student should feel no obligation to work with or even consider the possibility of working with individuals with whom the student does not feel he/she could get along. However, when the two parties involved make a commitment to work together, it must be considered a strong commitment.
Question Two:
When should the search for a thesis advisor start and what steps will usually occur?
Answer:
This usually takes place, or at least should take place, soon after notification of passing the Ph.D. qualifying exams. The first step after tentatively finding a thesis advisor is to form a supervisory committee. The members of the committee are selected to have expertise in areas which may be relevant to the student's thesis. Usually before intensive work on the thesis begins the student will complete the biology project requirement. The biology project is a separate requirement for each student and is set up with approval and guidance from the student's committee. The committee often uses other faculty to help administer the biology requirement.
Question Three:
How does a thesis progress? How often do student and thesis advisor meet? For what length of time do they meet?
Answer:
The specifics vary among thesis advisors and students. What follows are my preferences with regard to thesis advising. The initial period of time is spent seeking an appropriate question or topic to work upon. When the topic is decided upon, the student usually performs a literature search and summarizes work known to be done to that date. As soon as I take on a thesis student it is my preference to set up a regular appointment once a week for a length of time not to exceed one hour. (A moderate fraction of these meetings may be missed because of travel commitments.) By having a regularly scheduled meeting, the thesis advisor is able to have a good grasp of the student's progress and to have more input and feedback as the work progresses. Regularly scheduled meetings also allow the student to discuss problems that arise and to seek advice. In general, the art of thesis advising is trying to point the student by a few key directions, but to give as little detailed instruction as possible. For this reason students may become frustrated (as discussed below in Question Four).
In working on the topic the student will try to do something each week. This something may involve reading a paper and presenting a summary of the results back to the advisor, discussing the progress and/or lack of progress, attacking certain problems, discussing the topic, discussing the organization of data collection or Monte Carlo work, etc.
Experience has shown that it is desirable, although painful, to have the student write up the results of the material as findings are arrived at. It is not unusual, if this is not done, for the student to try six months later to reconstruct results that were "known" and not to be able to do so. By spreading the writing task out over the course of thesis work it makes the writing up of the final thesis, as well as the reading of the thesis, easier for both the student and the faculty member.
Question Four:
What problems are often encountered in writing a thesis?
Answer:
It is usual to hit "dry" periods while working on a Ph.D. thesis. One of the purposes of the Ph.D. thesis is to get the student involved in research. Of necessity, the progress cannot be predicted. It often goes quite slowly. I like to have my students read the Dover paperback book by Jacques Hadamard entitled The Psychology of Invention in the Mathematical Field. The purpose of reading this book is to see how research proceeds, to prepare for dry periods, and to recognize that by thinking about and working hard on a problem, even while not making progress, you may be triggering your subconscious mind to work on the problem. Because the Ph.D. thesis constitutes real research where the answer is not known going into the problem, it is expected that not only will there be dry periods, but occasionally it may necessary to switch problems. Particularly for individuals working upon difficult theoretical problems the problem considered may prove to be intractable if not impossible.
Question Five:
What is expected in a Biostatistics thesis? Must the topic be biologically oriented or is it o.k. to have a nonbiologic topic? Is it necessary to have theoretical and applied work or is it possible to have a thesis that is purely theoretical or purely applied? Can a thesis consist only of Monte Carlo work? Is it appropriate to have a thesis which consists only of data analysis on a particular project? How much material should be in a thesis? Is there some fixed length of time that is usual for the writing of a thesis?
Answer:
There is no statement from the Graduate Program about what constitutes a thesis. The thesis is to be independent work and a scholarly contribution to human knowledge. There are differing opinions among the Biostatistics faculty as to what constitutes an acceptable Ph.D. thesis. It is appropriate that the student discuss this with the thesis advisor before accepting an individual as a thesis advisor. Some individuals, myself included, feel that in general there should be at least some theoretical work as well as Monte Carlo work and/or data analyses. Other individuals feel that the creativity needed to design and program efficiently a Monte Carlo study may constitute a thesis. Still other individuals feel that the construction of a good mathematical model and its application to real data and/or Monte Carlo evaluation constitutes an acceptable thesis. Some individuals feel that data analyses on a very difficult project would make an excellent thesis. The opinions within the Graduate Program in Biostatistics differ from thesis advisor to thesis advisor. It is my personal opinion that it is a strength of the Graduate Program that a wide variety of projects might be appropriate. You should, however, consider what type of work you would like to do for a thesis before selecting an advisor and then to discuss, with potential advisors, what they feel constitutes an acceptable thesis.
In general, it is expected that the thesis topics have some relationship to mathematical or statistical techniques that can be used in biological fields. It is not necessary, however, for every thesis to explicitly have a biological component. In particular, theses that are more theoretical often deal with statistical and mathematical techniques that may be applied in a variety of settings, including some biological application.
It is the responsibility of the supervisory committee to oversee the course of the thesis work. In practice, however, it is usually the thesis advisor who determines when the student has adequate substance for a thesis. Undoubtedly the expectations of different faculty members as to the substance of material that constitutes a thesis differ considerably. (I personally do not have any data on this and find it hard even to make a general statement.) It is clear, however, that the substance of the thesis is much more important than the number of pages. A few brilliant results might be found in 15 or 20 pages and be a superior thesis. In a thesis that has more data analyses and/or Monte Carlo work, the length is usually considerably longer because of the necessity to present tables, graphs, etc. The important point to note is that the student is trying to make progress on a certain topic and not to produce a certain number of pages.
There is no "usual" length of time for writing your thesis. Progress in research necessarily varies greatly depending upon many factors.
Question Six:
What happens if there appears to be great difficulty in the progress of a thesis?
Answer:
It is best for the student to first, out of courtesy and also common sense, discuss the problems that you see with your thesis advisor and to let the advisor know your feelings. One of the important functions of a thesis advisor is to give psychological support to the student as the student enters what for many is a very uncertain period while they try their hand at statistical research. Nevertheless, it will occasionally occur that the student, for one reason or another, is not making much progress. In this case, after talking with the advisor it is possible to change thesis advisors. Either the student or the thesis advisor may suggest a change in this case. In general this could be done without great ill will on either side. Such a switch may be akward and psychologically difficult but can be made easily, although most students complete their work with the original advisor.
If the student feels that there are problems that need discussion and cannot be discussed with the thesis advisor, it is appropriate for the student to go to the Director of the Graduate Program in Biostatistics or the Chairperson for the department in which the thesis advisor teaches. It is a very rare event for things to get to this stage. In line with this the student should be aware that is not unusual to have "false starts" working on a Ph.D. thesis and for problems to change. (In my own experience the actual topic of my thesis was the fourth problem area that I attempted.)