Research Examination Guidelines and Policies
May 1, 2003; revised January 1, 2005; revised October 28, 2011
These guidelines and policies are intended to help Ph.D. students and advisers who are preparing for the research examination that is part of the revised Ph.D. program in CSE at UCSD. The guidelines are consistent with the policies stated in our Doctoral Program Overview and are an implementation of the principles explained in the Research Exam Overview. Sentences in italics below are taken from that overview. The guidelines here are quite specific, in order to promote consistency and fairness between students, and at the same time to establish and maintain high standards.
Policies and Procedures
All the rules here are mandatory. However, an adviser may petition for an exception or a change on behalf of a particular student.
Each petition must be in writing, must be sent to the chair of the CSE graduate committee and must come from the adviser (not the student). The petition must include a written explanation of how granting the special request to one student is nevertheless fair to all students.
A petition will be referred to the whole graduate committee if the committee chair so chooses, or if the adviser so requests. The decision of the whole graduate committee, by majority vote, will be final.
The graduate affairs office will keep a record of all petitions, both granted and denied, in order to help the graduate committee be fair in considering petitions.
The research exam is internal to the department. It is not the university qualifying examination (UQE), which is also called the thesis proposal exam in CSE. The research exam may not be merged with the thesis proposal exam.
A student who passes the research exam, who also meets the other M.S. requirements, and who does not already have an M.S. degree, is "eligible" to apply for the M.S. degree. A successful research exam counts as the Comprehensive examination requirement for Ph.D. students wanting to complete the M.S. degree along the way to completing the Ph.D. degree. If a student admitted to the Ph.D. program desires a terminal M.S. degree, and does not already have an M.S. degree, the student can take a Research Mastery examination as the M.S. Comprehensive examination. The Research Mastery examination has the same requirements as the Research exam, except where noted below.
The committee for the research exam consists of three members: the chair chosen by the graduate committee, and at least two other committee members from the CSE senate faculty. The student's advisor is not a member of the committtee, but is free to attend the research examination. The committee for the research mastery examination is the same, except that the second committee member is jointly chosen by the student and the committee chair rather than being the student's advisor.
The chair must be a tenured CSE faculty member, and at least one of the other members must be a tenure-track CSE faculty member. The third member may be anyone eligible to be the "sixth member" of a doctoral committee, as defined by OGSR. (For example, this includes adjunct professors and project scientists.)
With the consent of the committee, other faculty and researchers may participate in deliberations. Research exams are open to all.
Before attempting the research examination or research mastery examination the student must contact the staff graduate advisor at least TWO months in advance to submit a Ph.D. Research Exam Chair Request Form. At this time the student must complete a form specifying (1)whether the examination is a research examination or a research mastery examination, (2)the research advisor, (3)the proposed title/topic of the examination paper. The student returns the form to the staff graduate advisor. Significant departures from this initial information require the student to submit a revised form.
Upon receiving the form, the staff graduate adviser will then notify the graduate committee that the student is attempting the exam. At this point, the graduate committee appoints the chair of the student's committee. For the research examination, the chair discusses the scope of the examination with the student's advisor, and solicits suggestions for the other two committee members. For the research mastery examination, the chair discusses the scope of the examination with the student, and picks the two remaining members of the committee based on input from the student. The committee and the student determine the scope of the examination.
A student who fails on the first attempt at the research exam may retake it. A student who fails on the second attempt may petition the graduate committee for permission for a third attempt. The petition must be signed by the student's adviser, unless the student is proposing to take the research mastery examination.
A retake may not occur less than two months after the previous failure. The committee for a retake must have at least one member in common with the previous committee.
Students should aim to pass the research exam before the end of the sixth quarter in the Ph.D. program (not counting summer quarters). The normative time for passing the research exam is by the end of the second year of study. A petition to the CSE Graduate Committee is required to take the research examination after the student's seventh quarter of study. Students must pass the research exam before the end of the ninth quarter in the Ph.D. program (not counting summer quarters).
A student who fails on the third attempt, or who is denied a third attempt, or who exceeds the limit of nine quarters, must leave the Ph.D. program and is not eligible for financial aid beyond the end of the current quarter. (For this rule, the summer counts as a quarter.)
Before the oral part of the research exam, the student will produce a written project report. The report must be made available at least two weeks before the exam; any member of the faculty may see this report as soon as it is available.
A student who passes the oral exam must make any changes requested by the committee within two weeks, unless the committee sets a different deadline. The adviser will approve changes made by the student and forward the revised report for publication as described in the next paragraph.
A departmental web site will contain the written reports of [successful] past research exams. This web site will be accessible to CSE faculty, researchers, staff and Ph.D. students only. If a report contains unpublished material, the student and the committee may agree that the written report should be temporarily available to CSE faculty only. If so, the report will become public after a fixed period of time, by default one year.
The research examination is meant to assess a student's aptitude for conducting research. As such, the written report and presentation will be a combination of a creative survey and an optional original research component. The creative survey must provide deep insight into existing research on a specific topic. It must demonstrate the student's ability to read, understand, assimilate and synthesize existing research papers. The student should exhibit awareness of the research frontier in an area and preparedness to begin research in the area. The creative survey should also address weaknesses and gaps in the published work. Additionally, it can describe research programs to address these gaps. The optional original research component of the examination can cover original research that the student has conducted under their advisor's supervision. The research component can account for a significant portion of the examination, but it must still be accompanied by a substantial creative survey component. The student should develop their own written exposition and oral presentation of the research.
When writing the report, the student may receive advice and editorial assistance. However, the writing and content of the report must be the work of the student alone. S/he must qualify according to standard academic norms as the sole author of the report.
In an "acknowledgments" section at the end of the report, the student must disclose all assistance received, including from the adviser. The disclosure must include enough detail to make clear that the student qualifies as the sole author of the report. If a student receives undisclosed assistance, s/he is subject to penalties for academic dishonesty.
The examination is meant as a test of the student's technical communication skills as well as subject mastery. The research exam will verify the student's ability to communicate past and proposed future research, both orally and in writing. The quality of the writing and organization of the report must be close to that of papers published by good conferences in the area of the exam. The quality of writing and organization of the report for the research mastery examination must meet high standards for the internal communications of a corporation or other organization.
The research exam is intended to be a strong predictor of success in Ph.D. research. Students who do not produce a written report close in quality to a respectable conference paper will not pass the research exam.
To set a consistent and fair standard for the amount of work required, the length of the written report must be between 6000 and 9000 words, not counting figures and references. The report must be presented and formatted well, in a style similar to that prescribed by a good conference in the area.
The oral part of the research exam will be organized like a thesis proposal exam. First, the student will give a presentation lasting about 50 minutes. Second, the audience may ask questions. Third, the audience will leave; any faculty and the committee members may ask further questions. Fourth, the student will leave and the committee will deliberate. Last, the members of the committee will fill out an evaluation form.
The evaluation form is standardized and is designed to elicit feedback that is probing, clear, and useful. The committee must fill it out as a group and sign unanimously, in order to provide guidance to the student that is consistent. The completed evaluation form will become part of the student's permanent file. A copy will be given to the student.
For each item on the evaluation form, the committee must unanimously rate the student "excellent," "good," "passable," or "unsatisfactory."
- The rating "excellent" means at the level of a good paper or presentation at a top conference in the area of the exam. This rating is expected to be rare.
- The rating "good" means at the level of a paper or presentation acceptable by a respectable conference in the area of the exam.
- The rating "passable" means that improvement is necessary, but the committee is confident that the student will improve to the "good" level in less than a year.
- The rating "unsatisfactory" means below the "passable" level.
In order to pass the research exam, a student must receive a unanimous evaluation of "passable" or higher on every item on the evaluation form.
The chair ensures that all research exam policies are followed, maintains high standards, and leads the committee to a unanimous evaluation of the student. The chair and committee may utilize the collection of previous research examinations to compare the student's report to high quality reports from previous years. These reports can also be utilized to determine a suitable scope for the examination, in terms of the breadth of the area covered, and the quantity and difficulty of source materials surveyed.
The adviser provides guidance to help the student pass the research exam, and ensures that the student is genuinely the sole author of the written report.
Clarifications and Explanations
It is not intended that the student's report be publishable in a major conference, because the report need not contain original research. However, the quality of the writing must be close to the level seen at respectable conferences in the student's area. Similarly, the depth and breadth of the analysis in the report must be at least close to this level.
For any of the five questions, if a student is close to but not quite at the level of a respectable conference, s/he may still pass, if the committee assigns the grade "passable." The committee should use this grade if the student does not quite meet the "good" level, but s/he is making clear progress in the Ph.D. program and the committee has confidence that the student will reach the "good" level in the following year.
The four grading levels, with their specific descriptions, exist to make the evaluation of students less idiosyncratic and more fair, and also to provide clear feedback to each student. Grades will not be published but will be recorded in each student's file.
The details of what is expected for the research exam will vary from field to field. This is one reason why past reports will be public. Reports will be published on a web site that is accessible to CSE faculty, researchers, and graduate students only.
It is up to the committee whether or not a report must be based on original implementational work. Depending on the field, an insightful survey may or may not require software development and/or running experiments.
The length limit on research exam reports has three purposes: to speed progress through the research exam for good students, to make the workload reasonable for committee members, and to prevent a potential "arms race" between students writing ever longer papers. For illustration, the length limit corresponds to a paper in SIGMOD two-column format of length seven to ten pages.
The role of the committee chair is to promote fairness combined with high standards in the exam, in an impartial way. The adviser may not be the chair, because the adviser will often want to be an advocate for the student.
The adviser may help the student prepare for the research exam, and may act as an editor for the report. However, this assistance should consist of advice, not active co-writing. As Dean Tullsen has put it, the adviser can use a red pen, but not Emacs.
It is desirable for a student to pass the research exam within two years, but a student may take three years and still be in good standing. For example, a student may delay taking the research exam if he or she becomes involved in original research quickly.
Two months is the minimum time before a retake, not the maximum. This minimum gives the student and the adviser an incentive to make major improvements and to do genuinely well the second time.
The graduate committee has a transition plan for putting into practice the policy governing the new research exam. Initially, the chairs of research exam committees will be current members of the graduate committee. These faculty will meet to make sure they are setting fair and reasonable standards. The published reports from early research exams will provide guidance for future students, advisers, and members of research exam committees.