Department of Statistics
Ph.D in Statistics
The Ph.D. program prepares students for pursuing original research and the scholarship in academia, industry and government. Its primary purpose is to prepare professionals for research contributions to statistics and its applications. A distinguishing characteristic of the Ph.D. program is its research dissertation. The program aims to develop in students:
- A thorough knowledge of research methodology and its applications.
- The foundations necessary to understand new statistical theory tools as they develop.
- The capability for undertaking an original research investigation which will make a substantial contribution to the science of statistics.
- The ability to communicate statistical knowledge effectively.
- The ability to understand the relationships and interactions of statistics as a discipline with other fields.
Students in the Ph.D. program are required to complete a minimum of 48 semester hours of graduate course work beyond the bachelor degree or 18 semester hours beyond the MS degree from Temple. They must pass a Screening exam administered in June each year, a Proposal exam in area of dissertation research, accomplish the proposed dissertation research to write and finally defend it. Additionally, the students are expected to attend regularly scheduled department seminars and satisfy certain residency requirements set forth by the Graduate School.
The GAC will annually monitor the progress of students who have passed the Screening exam to ensure that no unjustified delays occur in completion of the degree requirements.
Required course work is not rigidly packaged to satisfy the minimum requirement of 48 semester hours (16 courses) beyond the baccalaureate degree. The 48 credit hours must include Stat 8001- 8002 (theory of statistics) and Stat 9001-9002 (advanced theory of statistics) and at least two more 9000-level courses not including the dissertation credits. Elective courses from disciplines other than Statistics may count towards the required course work with prior approval of the Program Director. The courses numbered 9000 or above offered outside of the Department of Statistics may not be recognized as advanced courses. This determination will be made by the Program director.
The students should recognize that completion of the 48 semester hours of course work may not necessarily be an adequate preparation to achieve the degree of proficiency demanded by the Proposal exam and the Dissertation research. They may expect to enroll in advanced topics seminars (9180 and 9190) as offered and/or engage intensive individual study.
Generally for 9000-level courses, 8000-level courses are prerequisites. Students with an M.S. degree from another university may have equivalent courses to our 8000-level prerequisites as part of their previous program. With the Program director’s approval, some courses from other institutions may be transferred within the limit set forth by the Department and the Graduate School’s policy.
All students in the Ph.D. program must pass a Screening Exam. That is,
- All full-time students in the Ph.D. program having completed Stat 8001 and 8003 in fall, and Stat 8002 and 8004 in spring of their first year will take the Screening exam in June.
- All part-time students in the Ph.D. program once they completed Stat 8001, 8002, Stat 8003 and 8004 will also take the Screening exam in June.
- In rare cases, students with MS in Statistics with a solid training in theoretical statistics from another institution who were granted waiver from taking Stat 8001, 8002, 8003 and 8004 courses by the Program director will also take the screening exam in June.
The Screening exam is closed-book and graded independently by three graders as Pass or Fail. Students failing the exam on the first try, may make take it again next year to pass it. Those failing twice will be advised to terminate the Ph.D. program with an option to receive the MS degree provided its requirements are met.
The Screening exam will cover the following topics in more general areas.
Probability: probability axioms, marginal, joint and conditional probabilities, Bayes’ theorem, random variables, marginal, joint and conditional distribution functions, expectation, moment generating function, Chebyschev’s theorem, Jensen inequality.
Common Probability distributions: bionomial, Poisson, hypergeometric, geometric, uniform, normal, exponential, Weibull, beta, gamma, multinomial, multivariate normal, conditional normal, covariance matrix, distributions of functions of random variables.
Sampling Distributions: chi-square, F, t distributions resulting from normally distributed random samples, distribution of order statistics
Asymptotics: convergence in probability, convergence in distribution, weak law of large numbers, the central limit theorem, delta method.
Estimation: unbiasedness, minimum variance, Cramer-Rao bound, sufficiency, completeness, Rao-Blackwell theorem, maximum likelihood, least squares, interval estimation, loss function, Bayes’ estimation.
Hypothesis Testing: uniformly most powerful test, Neyman-Pearson lemma, likelihood ratio test, Wald and score test, union-intersection and intersection-union principles, goodness of fit test, testing in contingency tables
Regression Analysis of Variance: dummy variables, screening of independent variables, multi collinearity, residuals, outliers and influential observations, crossed and nested factors, interactions, variance components, multifactor analysis, mixed effects models.
Generalized linear Models, logistic regression, and Chi squatted Tests for count data.
Aim and Intent of the Screening Exam
- To allow only those passing it to enroll in the theory sequence Stat 9001-9002.
- To give the Department an earliest opportunity to advise students to either continue or discontinue the program or pass it in the second attempt following year.
- To involve students in research who passed the exam in their second year having been assigned faculty mentors. It must be noted that a faculty mentor initially assigned may not ultimately serve as the dissertation advisor for their mentees. Choosing a dissertation advisor will remain a mutual decision between the student and the faculty member.
- Students who fail the Screening exam also on the second attempt will be advised to withdraw from the Ph.D. program. They may still be eligible for th M.S. degree.
Students who completed most of the course work must register for 1 credit in Stat 9994, “Proposal Exam Preparation.” in the semester of the Proposal exam to be taken. The proposal exam consists of a formal written proposal and its presentation by the student to the Advisory Committee in the presence of an independent representative of the GAC, and all interested faculty and students. Written Proposal will consist of a general introduction of the specialty area, outlines of the proposed research for dissertation, and a detailed review of the relevant literature to the proposed research. The Advisory Committee will evaluate if the proposed research is likely to lead to an original piece of research for the dissertation and student’s ability to complete the proposed research.
The following guidelines must be followed for completing the Proposal exam.
- Advisory Committee and the GAC should have a written research Proposal at least three weeks in advance of scheduling the Proposal exam. Also, a copy of the Proposal must be placed in the Reading Room clearly marked “For the Reading Room only”.
- The date and time of the Proposal exam must be set with the approval of the advisory committee and the GAC.
- The Proposal exam must accompany a separate manuscript consisting of some accomplished research being proposed. It should follow guidelines of the journal to which it is being submitted.
- The manuscript may have coauthors but the Proposal itself should have only the student’s name.
- The research Proposal need not be too lengthy a document. It should clearly outline the research problem, must contain a review of the pertinent literature focused in the proposed research area followed by some accomplished research towards dissertation, and finally the references.
In addition to the Department of Statistics’ guidelines, students and advisors must also adhere to those of the Graduate School.
Having passed the Screening and the Proposal exams, student will have attained the status of the doctoral candidate after filing the Dissertation Proposal Transmittal Form to the Graduate School and approved written Proposal. Doctoral students must maintain continuous enrollment remaining enrolled in Stat 9999.
Dissertation Research Defense
Doctoral candidates must register each fall and spring semester, and in the semester in which the defense is held, for Stat 9999, “Dissertation Research.” The Graduate School requires Doctoral Examinations/Culminating Experiences to have a minimum of 6 semester hours with at least 2 of them required to be in course number 9999. The remaining 4 semester hours can be a combination of the following course numbers: 9994, 9998, and/or 9999. Upon completion of the proposed research leading to Ph.D. dissertation, the doctoral student will present a formal oral defense to the Advisory Committee, and interested faculty and students..
At the defense, student must have at least one manuscript formally prepared from the dissertation research for submission to a professional journal, if one has not been already submitted. It is to ensure that at least a portion of dissertation research is published in a refereed journal.
The steps leading to the final defense are as follows:
- The doctoral candidate will submit a final draft to the Advisory Committee. The Committee will have three weeks to approve or disapprove it or suggest modifications.
- Taking into account the comments of the committee, students will revise the dissertation draft for distributing to the Advisory Committee, representative of the Graduate School, and to the GAC at least three weeks before scheduling the defense
- Attendant at the defense must be members of the dissertation advisory committee, representative of the GAC if not already a member of the advisory committee, and a faculty representative of the Graduate School from outside of the Department.
- The date and time of the defense must be mutually convenient to all concerned. The defense will be open to all University faculty and students.
Students in the Ph.D. program are expected to attend regularly held department seminars. Students’ attendance at seminars will be recorded. The program director may prevent students found in serious violation of this departmental expectation from taking the Proposal and the defense of dissertation exams.
In the course of studies students will receive advising on multiple occasions from the Program Director. Soon after passing the Screening exam, students will be assigned faculty mentors for research activities. However, a faculty mentor may not ultimately serve as the dissertation advisor for their mentees. Choosing a dissertation advisor will remain a mutual decision between the student and the faculty member.
Having completed the core courses Stat 9001 and 9002, if not before, students should choose dissertation advisors from the faculty ranks of the Department of Statistics. In consultations with their dissertation advisors, students will form a dissertation advisory committee consisting of the faculty from within and possibly from outside of the Department.
Formal authority in the doctoral program resides with the Program Director. In addition, should a student decide to change dissertation advisor, it should be discussed first with the Program Director prior to making changes. The Program Director will ensure protection of academic rights of every one involved in the change.
Prior work in statistics is not a requirement for admission to graduate study. Applicants for admission to the graduate program must hold a bachelor’s or equivalent degree from an accredited college or university and must have a minimum undergraduate grade point average of 2.8 overall and 3.2 in mathematical courses. Applicants must submit scores on the aptitude portion of the Graduate Record Exam (GRE). The scores on each of the three components of the GRE (verbal, quantitative, and analytical) typically fall within the following percentile ranges:
- Verbal: 50 to 99
- Quantitative: 80 to 99
- Analytical: 50 to 99
Applicants with an undergraduate GPA of 3.4 or better in mathematical courses may substitute scores on the Graduate Management Admission Test (GMAT) for the GRE. The GMAT scores must also fall within the above percentile ranges. An undergraduate GPA of 3.4 or better in mathematical courses can be used to offset a marginal score on the quantitative component of the GRE.
Grades of B or better in college-level English courses can be used to offset marginal scores on the verbal and analytical components of the GRE. Foreign students whose native language is not English must submit scores on the TOEFL exam. The minimum score is 575, and the typical range is 590 to 660. A score above 610 may be used to offset a marginal score on the verbal component of the GRE.
Two letters of recommendation and a statement of goals are required. Admission is competitive. The applicant’s academic record and test scores are given the highest weight in the admissions process, although other factors, such as the applicant’s ability and potential for further study as demonstrated by research publications, are also taken into consideration. A personal interview is not required but highly recommended.
For more details contact: