PhD Requirements

Goals

Students completing the Ph.D. degree will be well-prepared for a job in industry, government, or at a university. Throughout the program, students will develop the following skills:

communication skills: e.g., explaining or reporting statistical results to non-statisticians, such as managers;
broad base of knowledge of modern statistical methods, including computing; ability to choose appropriate analyses when presented with a real problem and to understand the implications of these analyses (assumptions, shortcomings, interpretation of results);
ability to learn after finishing degree: i.e., to learn new statistical techniques from textbooks and journal articles, modifying the techniques, if necessary;
the ability to teach an introductory statistics class;
the ability to communicate to research statisticians both verbally and in writing, such as by giving papers/posters at conferences;
mathematical maturity;
creativity and originality: ability and potential to generate and solve "good" research problems;
deep knowledge of a speciality;
library skills, grant writing skills.

Time line

Full load of courses in the first year, followed by appropriate courses in subsequent years to help research and to broaden knowledge.
PhD reading course in first year; this is the current Qualifying exam.
Projects with faculty members to learn about research.
PhD Comprehensive exam before the end of the second year of study.
PhD thesis consisting of independent original research.
Presentation at the Department Seminar, and PhD defense (3 to 5 years to completion is expected).

Coursework

Students develop these skills through a variety of activities. Courses typically combine traditional lecture-based learning with independent learning, where students read relevant journal articles, download software from the web, analyze difficult data sets, and present and discuss their findings with their classmates. New statistical problems and approaches are discussed in the weekly statistics seminar (a research-level seminar), and in the research seminars and journal discussions of the Biostatistical Research Group. Many students serve as research assistants, often working directly with researchers in other areas such as medicine. Students also develop the above skills while working as Teaching Assistants and as lecturers for courses. And of course, research skills are developed while carrying out work for the Ph.D. thesis.

In addition to STAT 550 and 551, students should take a sufficient number of courses to insure that they have a broad exposure to statistics at the graduate level. The formal breadth requirement is that students take at least 15 credits of graduate level coursework (in addition to STAT 550 and 551 if necessary, and in addition to the PhD reading course) before being admitted to Candidacy (click here for more information). However, in cases where students have taken `advanced topics' coursework as part of their M.Sc. work, this requirement can be dropped to 12 credits or 9 credits, at the discretion of the Graduate Advisor. In many instances it may be academically wise to take more than the minimum number of courses. Also, students are encouraged to consider graduate courses in other departments where statistical methods play an important role. Such courses do count toward the breadth requirement.

PhD Thesis

A PhD thesis consists of original research, conducted independently by the student. The student can choose their own thesis topic and find an appropriate supervisory committee, or decide on a topic after choosing a supervisor. Faculty members generally do not pose specific problems for PhD research, other than perhaps to help the student get started on something concrete; in this case, the student is expected to independently develop the remainder of the thesis.

Deciding on a PhD topic is not easy; the ideal topic is something one is very interested in and wants to spend all of one's time researching. The PhD student should be thinking about potential thesis topics at the beginning of the PhD program. The student should familiarize himself/herself with the research of faculty members as soon as possible in order to choose a supervisor and supervisory committee. Good ideas are to take graduate courses from various faculty members in the first year and to talk to individual faculty members.

Doing PhD research also is not easy, since original research is quite different from coursework where the homework typically consists of "solved" problems. One way to learn about research is to talk to and possibly work with several faculty members on their current research projects. By working on these projects, you may learn about mathematical, statistical and computing tools not seen in coursework. If you are helpful on these projects, then benefits would research assistantships and co-authored research publications.

Qualifying exam

The PhD Qualifying Exam requirement has been changed. Previously, students wrote a `sit-down' exam in May of their first year. Effective for students entering the PhD program in 2005 or later, the new requirement is the successful completion of a "reading and research" course in the first year of the program. This course involves reading and reporting on a number of research articles in the statistics literature.In 2008-2009 this course will be offered as STAT 548 (directed studies).

Comprehensive exam

Normally, within the first two years of entering the program, the student writes a research proposal and presents it in the Comprehensive Exam.

The Comprehensive "exam" seeks to determine if the PhD student has the ability needed to do research well. That ability includes the technical skills developed from course work. However, it also assesses a variety of other prerequisites to research success and it is not expected that all PhD students possess these skills to the degree necessary for admission to candidacy.

Before reaching the Comprehensive exam, the student should have a pretty good idea whether they can do and enjoy doing independent research. If not, a research career can be very unsatisfying and success would be hard to achieve. Good modern research environments require the publication of the equivalent of a couple of good papers a year, year-after-year, while your progress is continuously monitored for promotion, grants and so on. Not something that easy to do without the requisite skills, including a strong motivation to do research.

More details here

Supervisory Committee

A thesis supervisory committee consisting of at least three faculty members should be formed when a topic is decided upon. A student should form a supervisory committee as early as possible but, at the very latest, this should be done promptly following the successful completion of the comprehensive examination.

The Departmental Graduate Committee should be informed of the composition of the supervisory committee once it is formed, and of any subsequent changes to the supervisory committee.

Defense and completion of program

PhD students are required to make a presentation based on their completed work in a Department Seminar. This requirement must be fulfilled before submission of the thesis for external review. At the conclusion of the Ph.D. program, the student must defend a written thesis at an oral examination. Throughout the program, students are encouraged to write up their research for submission to statistical journals. Note that having a publication record is a great asset when applying for an academic position during the final year of PhD studies.