Research Preliminary Exam
The goal of the Research Prelim is to evaluate a student’s readiness for PhD research in Statistics. To this end each student will read an approved chosen paper from the Statistical literature and will present an in-depth study of it, both as a talk and as a written report. Students must show a thorough understanding of the methods described in the paper and should reproduce key simulation and/or data analyses presented in the paper. There should also be “something added”, which could be, for example, a discussion of where the paper fits in the broader literature, or an additional related result or analysis.
The research prelim is normally based on a paper studied in one of the Spring preparatory classes in the three areas of Methodology, Stochastic Modeling, and Statistical Machine Learning. These classes are restricted to PhD students and aim to help students develop skills in reading, understanding and critiquing contributions to the literature. However, the class is not the prelim.
Timing of the Exam
The exact timing will depend on the availability of faculty examiners, and will vary by the subject area in which students take the exam. However, the written report will be due no earlier than the last day of Spring classes. The oral presentations should normally take place either in exams week or in the break week following Spring Quarter. Results of the examination will be available to students by early July.
Format of the Written Report
The report should be written in the style of a journal paper and should be no more than 20 pages, double-spaced, including all key Figures and Tables. It should have sections such as Introduction or Background, Methods or Model, Implementation or Results, and Conclusion or Discussion. It should have a Bibliography, and papers should be cited in a standard style. Appendices including Supplementary Material may also be submitted, but these will not necessarily be read by the faculty examiners. An important part of writing papers is to learn to present essentials fully and clearly yet concisely. Note that these requirements may differ from those for final reports written in the preparatory classes: the class is not the prelim.
Format of the Oral Presentation
Oral presentations will normally be scheduled at one-hour intervals. The first part of the exam is open to graduate students and faculty of the Department of Statistics. In this part, the student should expect to present for 25 to 30 minutes; with questions following, this open portion of the exam will total 30 to 35 minutes. A closed session normally of 10 to 15 minutes will follow in which the faculty examiners will further discuss the exam with the student. A brief period is allowed between exams, for the faculty to further discuss the exam outcome. Note that the oral presentation will generally differ from a final presentation given in the preparatory classes: the class is not the prelim.