The Physics Candidacy Examination
advancement to candidacy
Before embarking formally on her/his PhD thesis research, a student must become a candidate for the doctorate. The chief requirement for advancement to candidacy is to pass a comprenhensive written test administered by the physics faculty: the candidacy exam. The current structure of the exam was decided by a policy adopted by the Physics faculty on May 3, 2002, stated below. The purpose of the exam is twofold.
First, it assures that the
student has the requisite knowledge to undertake independent research at the PhD level. Second, it assures that
the student has the broad competency in the physics discipline that a doctoral degree implies. The exam poses problems in which the student must use physics reasoning to predict the outcome of a real or hypothetical measurement or observation. The exam tests for mastery of physics course material at the undergraduate level. The questions test for deeper and more integrated understanding than is normally required in an undergraduate course. A question often requires knowledge from more than one undergraduate course. Past exams on file in the library show the style and the level of the questions.
subject coverage
The central topics to be covered by the exam are Quantum Mechanics, Electromagnetism, and Classical Mechanics. It is to be expected that the exam also will regularly cover Statistical Mechanics/Thermodynamics and topics like Special Relativity. Questions may touch on more specialized topics, such as Mathematical Physics, Astrophysics, Solid State/Condensed Matter, Optics, Atomic Physics, Experimental Physics, High Energy Physics, and Short Answer Estimation. But such questions are to be within the scope of undergraduate material that the faculty takes to be indispensible.
studying for the exam
Since most students have not taken a broadly based exam like this before, they generally need to prepare for it. Students normally spend two months or more studying for the exam, reviewing their undergraduate course work and solving previously-posed exam problems for themselves. Often students study in pairs or groups.
Many faculty and senior students look back on their study for such exams as one of the most valuble parts of their PhD training.
taking the exam
Students who intend to take the exam are asked to inform the department office a couple of months before it is given. The exam is given in two six-hour sessions, usually on the third floor of Kersten. Students may bring a calculator capable of performing arithmetic and calculating standard mathematical functions, but not one which stores information to be consulted during the exam. The students are not allowed to consult outside information except for two items. A student may prepare a single sheet of standard sized office paper with his/her own handwritten notes. Students may consult a sheet provided at the exam listing physical constants and conversion factors, . Students are trusted not to use outside information beyond these sources. Students are free to leave the room. A member of the exam committee is reachable to answer questions that arise. Most students bring food and something to drink.
how the exam is written and graded
A committee of four faculty members is responsible for making up the exam. They are chosen to span the breadth of expertise in the department. It is administered just before the beginning of the autumn quarter and just before the beginning of the spring quarter. A number of measures are taken to assure the impartiality and accuracy of the exam. The questions are submitted by the faculty at large. Each on the exam question is revised by a committee member and is worked and critiqued independently by another member, to assure that the question is well posed and appropriate. The committee grades the exams without knowing who wrote them. Two committee members grade each exam question independently, without writing on the exam. If the two scores on a question differ by more than two points out of ten, those members review their notes, discuss that student's answer and resolve the discrepancy. Once the scores have been established, the committee meets to decide which exams deserve a passing grade. They decide what level of performance should be sufficient to pass. Historically this has corresponded to a threshold score of about half of the maximum score. In borderline cases, the committee looks at the student's performance on individual questions. They look for evidence of breadth and of solid reasoning.
Scores deemed insufficient for advancement to candidacy may nevertheless be high enough to qualify for a master's degree. Such students are given so-called M-level passes but are not admitted to candidacy.
announcing the results
Students are informed about whether they passed as soon as possible after the exam is graded. Solutions are made available later as an aid for future students. The committee doesn't attempt to explain its grading of individual exams to the students.
The Department will continue to administer and require that all students pass a Candidacy Exam.
- It will be a 2-day, 6-hour per day written exam.
- There will be no oral component.
- The exam will be offered twice a year at the same level of difficulty:
i) shortly before Orientation Week in September before school starts, and
ii) at the start of classes in the Spring Quarter (late March).
- All students must attempt the exam by the Fall of their second year.
- Students must pass the exam by the Fall of their third year.
- The exam is open to first-year students on entering. It is to be emphasized that performance on the exam, if taken on entering, will not negatively affect any student. Students who would prefer to wait until Spring or the following Autumn after entering are welcome to do so.
- The level of the exam will be that of advanced undergraduate courses. Problems will be selected so as to emphasize those parts of undergraduate physics which we take to be indispensable.
- There are only three levels of achievement on the candidacy --
Pass to PhD candidacy, Pass at the Masters' level, and Fail --
T. Witten June 16, 2003