The Statistics PhD Program

1. Overview

The Statistics PhD program is rigorous, yet welcoming to students with interdisciplinary interests and different levels of preparation. The program requires four semesters of residence.

Normal progress entails:

Year 1. Perform satisfactorily in preliminary course work. In the summer, students are required to embark on a short-term research project, internship, graduate student instructorship, reading course, or on another research activity.
Years 2-3. Continue course work. Find a thesis advisor and an area for the oral qualifying exam. Formally choose a chair for qualifying exam committee, who will also serve as faculty mentor separate from the thesis advisor.  Pass the oral qualifying exam and advance to candidacy during the spring semester of Year 2 or the fall semester of Year 3. Present research at BSTARS each year.
Years 4-5. Finish the thesis, and give a lecture based on it in a department seminar.

Relevant Forms:
Annual Progress Review
Selection of Thesis Advisor
Selection of Qualifying Examination Chair
Graduate Student Petition
Application for the Qualifying Exam (via CalCentral)
Application for Candidacy (via CalCentral)

2. Course work and evaluation

2.1 Preliminary stage: The first year

Effective Fall 2019, students are expected to take four semester-long courses for a letter grade during their first year which should be selected from the core first-year PhD courses offered in the department: Probability (204/205A, 205B,), Theoretical Statistics (210A, 210B), and Applied Statistics (215A, 215B). These requirements can be altered by a member of the PhD Program Committee (in consultation with the faculty mentor and by submitting a graduate student petition) in the following cases:

  • Students primarily focused on probability will be allowed to substitute one semester of the four required semester-long courses with an appropriate course from outside the department.
  • Students may request to postpone one semester of the core PhD courses and complete it in the second year, in which case they must take a relevant graduate course in their first year in its place. In all cases, students must complete the first year requirements in their second year as well as maintain the overall expectations of second year coursework, described below. Some examples in which such a request might be approved are described in the course guidance below.
  • Students arriving with advanced standing, having completed equivalent coursework at another institution prior to joining the program, may be allowed to take other relevant graduate courses at UC Berkeley to satisfy some or all of the first year requirements

2.2 Requirements on course work beyond the first year
Students are required to take five additional graduate courses beyond the four required in the first year, resulting in a total of nine graduate courses required for completion of their PhD. In their second year, students are required to take three graduate courses, at least two of them from the department offerings, and in their third year, they are required to take at least two graduate courses. Students are allowed to change the timing of these five courses with approval of their faculty mentor. Courses that are research credits, directed study, or departmental seminars do not satisfy this requirement (for courses offered by the statistics department any course numbered 204-272 satisfies the requirement for a graduate course).

Of the nine required graduate courses, students are required to take for credit a total of 24 semester hours of courses in the department numbered 204-272 inclusive. A member of the PhD Program Committee (in consultation with the faculty mentor and by submitted a graduate student petition) may consent to substitute courses at a comparable level in other disciplines for some of these departmental graduate courses. In addition, the committee member may waive part of this unit requirement.

2.3 Evaluation
First year course work: For the purposes of satisfactory progression in the first year, grades in the core PhD courses are evaluated as:
A+: Excellent performance in PhD program
A: Good performance in PhD program
A-: Satisfactory performance
B+: Performance marginal, needs improvement
B: Unsatisfactory performance

First year and beyond: At the end of each year, students must meet with his or her faculty mentor to review their progress and assess whether the student is meeting expected milestones. The result of this meeting should be the completion of the student’s annual review form, signed by the mentor (available here). If the student has a thesis advisor, the thesis advisor must also sign the annual review form.

3. Guidance on choosing course work

3.1 For students primarily interested in Statistics

Choice of courses in the first year: Students enrolling in the of fall 2019 or later are required to take four semesters of the core PhD courses in their first year. Students primarily interested in statistics would normally take the 210A,B sequence (Theoretical Statistics) and then one of the 205A,B sequence (Probability) or the 215A,B sequence (Applied Statistics), based on their interests (though students are allowed to mix and match, where feasible). Students who opt for 210AB and 215AB in the first year should be aware that 210B requires some graduate-level probability concepts that are normally introduced in 205A (or 204).

Postponement of one semester to the second year: As noted in the official course requirements, the PhD committee may, in some cases, allow for the postponement of one semester of the core courses to the second year. Students who would prefer to take 210A and 215A in their first year and are concerned about their preparation for 210B should discuss this issue with their faculty mentor; the faculty mentor may recommend that the student be allowed to postpone fulfilling one semester of their first year requirement to the second year, and substitute an alternative graduate level course for 210B in the spring. (Alternatively, the student could choose to take 204 in their first year instead of 210B in years in which 204 is offered in the spring). Such a student would then need to take either 204, 205A or 215B in their second year to complete the first year requirements. Similarly, for students with strong interests in another discipline, the faculty mentor might recommend that the student be allowed to postpone one of the spring core PhD courses to the second year.  Students who are less mathematically prepared might also be allowed to take the upper division (under-graduate) courses Math 104 and/or 105 in their first year in preparation for 205A and/or 210B in their second year. In consultation with the faculty mentor, a graduate student petition should be submitted to the PhD Committee for postponement and should generally be limited to only one semester. 

After the first year: Students with interests primarily in statistics are expected to take at least one semester of each of the core PhD sequences during their studies. Therefore at least one semester (if not both semesters) of the remaining core sequence would normally be completed during the second year. The remaining curriculum for the second and third years would be filled out with further graduate courses in Statistics and with courses from other departments. Students are expected to acquire some experience and proficiency in computing. Students are also expected to attend at least one departmental seminar per week. The precise program of study will be decided in consultation with the student’s faculty mentor.

Remark. Stat 204 is a graduate level probability course that is an alternative to 205AB series that covers probability concepts most commonly found in the applications of probability. It is not taught all years, but does fulfill the requirements of the first year core PhD courses. Students taking Stat 204, who wish to continue in Stat 205B, can do so (after obtaining the approval of the 205B instructor), by taking an intensive one month reading course over winter break.

Designated Emphasis: Students with a Designated Emphasis in Computational and Genomic Biology or Designated Emphasis in Computational and Data Science and Engineering should, like other statistics students, acquire a firm foundation in statistics and probability, with a program of study similar to those above. These programs have additional requirements as well. Interested students should consult with the graduate advisor of these programs. 

3.2 Probability

Starting in the Fall of 2019, PhD students are required in their first year to take four semesters of the core PhD courses. Students intending to specialize in Probability, however, have the option to substitute an advanced mathematics class for one of these four courses. Such students will thus be required to take Stat 205A/B in the first year,  at least one of Stat 210A/B or Stat 215A/B in the first year, in addition to an advanced mathematics course. This substitute course will be selected in consultation with their faculty mentor, with some possible courses suggested below. Students arriving with advanced coursework equivalent to that of 205AB can obtain permission to substitute in other advanced probability and mathematics coursework during their first year, and should consult with the PhD committee for such a waiver.

During their second and third years, students with a probability focus are expected to take advanced probability courses (e.g., Stat 206 and Stat 260) to fulfill the coursework requirements that follow the first year. Students are also expected to attend at least one departmental seminar per week, usually the probability seminar. If they are not sufficiently familiar with measure theory and functional analysis, then they should take one or both of Math 202A and Math 202B. Other recommended courses from the department of Mathematics or EECS include:

  • Math 204, 222 (ODE, PDE)
  • Math 205 (Complex Analysis)
  • Math 258 (Classical harmonic analysis)
  • EE 229 (Information Theory and Coding)
  • CS 271 (Randomness and computation)

4. The Qualifying Examination 

The oral qualifying examination is meant to determine whether the student is ready to enter the research phase of graduate studies. It consists of a 50-minute lecture by the student on a topic selected jointly by the student and the thesis advisor. The examination committee consists of at least four faculty members approved by the Graduate Division, three from the Statistics Department and one Academic Senate Faculty member from another department.

Qualifying Exam Chair. For qualifying exam committees formed in the Fall of 2019 or later, the qualifying exam chair will also serve as the student’s departmental mentor, unless a student already has two thesis advisors. The student must select a qualifying exam chair and obtain their agreement to serve as their qualifying exam chair and faculty mentor. The student's prospective thesis advisor cannot chair the examination committee. Selection of the chair can be done well in advance of the qualifying exam and the rest of the qualifying committee, and because the qualifying exam chair also serves as the student’s departmental mentor is expected to be selected by the beginning of the third year or at the beginning of the semester of the qualifying exam, whichever comes earlier. See Advising below for more details regarding the selection of the Qualifying Exam Chair.

Paperwork and Application. Students at the point of taking a qualifying exam are assumed to have already found a thesis advisor and to should have already submitted the internal departmental form to the Graduate Student Services Advisor (found here).  Selection of a qualifying exam chair requires that the faculty member formally agree by signing the internal department form (found here) and the student must submit this form to the Graduate Student Services Advisor.  In order to apply to take the exam, the student must submit the Application for the Qualifying Exam via CalCentral at least three weeks prior to the exam. If the student passes the exam, he or she can then officially advance to candidacy for the Ph.D. If the student fails the exam, the committee may vote to allow a second attempt. Regulations of the Graduate Division permit at most two attempts to pass the oral qualifying exam. After passing the exam, the student must submit the Application for Candidacy via CalCentral.

4. The Doctoral Thesis

The Ph.D. degree is granted upon completion of an original thesis acceptable to a committee of two departmental faculty and an outside member. The thesis should be presented at an appropriate seminar in the department prior to filing with the Dean of the Graduate Division. See Alumni if you would like to view thesis titles of former PhD Students.

Graduate Division offers various resources, including a workshop, on how to write a thesis, from beginning to end. Requirements for the format of the thesis are rather strict. For workshop dates and guidelines for submitting a dissertation, visit the Graduate Division website.

6. Teaching Requirement

For students enrolled in the graduate program before Fall 2016, students are required to serve as a Graduate Student Instructor (GSI) for a minimum of 20 hours (equivalent to a 50% GSI appointment) during a regular academic semester by the end of their third year in the program.

Effective with the Fall 2016 entering class, students are required to serve as a GSI for a minimum of two regular academic semesters and complete at least 40 hours prior to graduation (20 hours is equivalent to a 50% GSI appointment for a semester) for Statistics courses numbered 150 and above. 

7. BSTARS

Each spring, the department hosts an annual conference called BSTARS. Both students and industry alliance partners present research in the form of posters and lightning talks. All students in their second year and beyond are required to present a poster at BSTARS each year. This requirement is intended to acclimate students to presenting their research and allow the department generally to see the fruits of their research. It is also an opportunity for less advanced students to see examples of research of more senior students. However, any students who do not yet have research to present can be exempted at the request of their thesis advisor (or their faculty mentors if an advisor has not yet been determined).

8. Mentoring for PhD Students

Initial Mentoring: PhD students will be assigned a faculty mentor in the summer before their first year. This faculty mentor at this stage is not expected to be the student’s PhD advisor nor even have research interests that closely align with the student. The job of this faculty mentor is primarily to advise the student on how to find a thesis advisor and in selecting appropriate courses, as well as other degree-related topics such as applying for fellowships.  Students should meet with their faculty mentors twice a semester. This faculty member will be the designated faculty mentor for the student during roughly their first two years, at which point students will find a qualifying exam chair who will take over the role of mentoring the student.

Research-focused mentoring: Once students have found a thesis advisor, that person will naturally be the faculty member most directly overseeing the student’s progression. However, students will also choose an additional faculty member to serve as a the chair of their qualifying exam and who will also serve as a faculty mentor for the student and as a member of his/her thesis committee. (For students who have two thesis advisors, however, there is not an additional faculty mentor, and the quals chair does NOT serve as the faculty mentor).

The student will be responsible for identifying and asking a faculty member to be the chair of his/her quals committee. Students should determine their qualifying exam chair either at the beginning of the semester of the qualifying exam or in the fall semester of the third year, whichever is earlier. Students are expected to have narrowed in on a thesis advisor and research topic by the fall semester of their third year (and may have already taken qualifying exams), but in the case where this has not happened, such students should find a quals chair as soon as feasible afterward to serve as faculty mentor.

If there is a need for a substitute faculty mentor (e.g. existing faculty mentor is on sabbatical or there has been a significant shift in research direction), the student should bring this to the attention of the PhD Committee for assistance.

Student matters:

A lot of information is also available from our very active Statistics Graduate Students Association (SGSA):

You can also find comments from former students here.

Last revised March 2019