Upon completion of the necessary courses (Statistical Models & Methods I, II, III and Statistical Inference I, II) the student will be given a written preliminary examination, usually in January of the second academic year. This examination will be organized and administered by the graduate studies committee. The exam will consist of two parts - Applied Statistics and Theory of Statistics. The applied part will cover Statistical Models and Methods I, II and III, Clinical Trials, and Biostatistical Computing and Data Management. The theory part will cover the materials from Statistical Inference I & II. This will be a standard divisional exam, and evaluation will be done by the whole faculty. The criteria for evaluation will be based on student's understanding and competency in basic principles and foundations of biostatistics, and his/her potential for conducting independent research in statistical methods and applications. If a student does not pass this exam, he/she will have a second opportunity to take it. The preliminary examination will be offered every January and July by the Division. The student must pass this examination to continue in the PhD program.
The student is required to take BIOST 295 Readings & Research for 3 credit hours each with two different members of the faculty. Typically this is done in the first two summers and in the process of selecting a dissertation topic and advisor.
Upon successful completion of the preliminary exam and the required biostatistics courses (usually at the end of the third year), the student will be given a qualifying examination. This examination is tailor-made for each student, and it is organized, administered and evaluated by his/her advisory committee. The evaluations will be based on student's in-depth understanding and competency in advanced topics in biostatistics, and his/her ability and maturity to apply the knowledge earned from the course-work in doing meaningful research. The exam consists of two parts. The first part will be an oral examination testing the student's general statistical knowledge at the advanced level. The second part consists of writing a dissertation proposal and presenting it to the division. This proposal must be approved by his/her advisory committee. A student not passing either part of the exam may be given another chance to retake that part within three months of the first attempt. Students passing this exam will be admitted to PhD candidacy.
The PhD candidate must submit a dissertation representing an original research contribution. It must show high attainment and clear ability to carry out independent biostatistics research of publishable quality. The final oral examination will be administered by his/her advisory committee after the student has completed all other formal requirements for the PhD degree. It will be a public defense of the dissertation. The student also will be expected to demonstrate a good understanding of materials relevant to the general field in which the dissertation is written. The student's advisory committee will evaluate the performance of the student in the dissertation defense.
The student begins his/her dissertation research during the third year. The initial step consists of identifying a topic that is of mutual interest to the student and a member of the faculty who serves as the dissertation advisor. Courses, talks and presentations by the faculty assist the student in this process. After a literature survey and a clearer definition of the scope of the research under the direction of the advisor, the student submits a written proposal and presents it orally to the advisory committee. During the conduct of the dissertation research the advisory committee meets periodically to monitor the student's progress. Upon completion of the proposed research the student submits the dissertation and defends it in a public presentation.
The dissertation must be an original contribution to scientific knowledge. It can involve development of new statistical methodologies, evaluation of existing methodologies and study of their properties, innovative application of existing methodologies, or any combination of the above. The dissertation should be of publishable quality in peer reviewed journals in biostatistics or statistics.