I was going over this post by Incoherent Ponderer (IP) on the subject of graduate admissions, as well as some others to which he links. I agree with all of the very sensible remarks that IP makes, but I wish to address here one specific issue a bit more concretely. It is a dilemma that graduate admission committees (GACs) often face, when trying to decide to which applicants to extend an offer.
I am talking here about admission into a science graduate program whose main focus is research, and whose culminating point is the completion of an original piece of research work, described in a dissertation written by the student. This is clearly the case for any doctoral program, but applies to some Master’s degree programs as well. Because of the emphasis on research, GACs must obviously base their decisions on the predicted ability of an applicant to excel in research. Graduate course work should be seen as propaedeutic to research, i.e., it is only important to the extent to which it can impact a graduate student’s research performance.
There are, of course, some cases that are clear-cut (one way or another), but generally few of them can be classified as such. More often, one has to make difficult choices, and time and again the GAC finds itself in the situation of having to pick one of two candidates more or less fitting the following profiles:
— Candidate A has a strong academic record, high grade point average, good GRE score, but little or no research experience. Candidate’s statement expresses interest for the scientific discipline and enthusiasm for research, perhaps openness toward different areas of inquiry, but fails to mention a specific one in which (s)he wishes to carry out graduate work, much less identify a faculty in the department hosting the program, as a potential supervisor.
— Candidate B has an academic record that is good but not stellar, or sometimes even on the weak side, as well as some hands-on research experience as an undergraduate student. (S)he expresses a clear interest for a specific area of research, typically the one in which undergraduate work was carried out, and/or names one or more program faculty as possible supervisor(s) .
Which applicant should be given priority, if a choice between one of the two has to be made ?
My experience is that GACs, when facing such a quandary, will typically go for B, the main motivations being:
1) The program is ultimately about research, not course work, and “students who are academically strong do not necessarily make the best researchers”. Previous research experience is almost always taken a reliable indicator of talent and inclination for research work.
2) Why admit into the program a student who is unlikely to find a supervisor, as (s)he has ostensibly no well-defined research interests ? Does such an applicant not obviously lack focus and/or motivation ?
Now, while I do not feel comfortable stating that I would always give preference to A instead, my experience has made me very wary of the above motivations. I regard both of them as dubious, and I have seen how they can lead to the exclusion of strong applicants, to the benefit of others who do not work out in the end. Here is why:
1) That academically strong students do not always make the best researchers is true — but from that to claiming or implying that they never do, there is a bit of a leap. In fact, the contention that academic record has nothing to do with research ability seems bizarre to me.
My own observation is that there is in fact a very strong correlation between excellence in research and a distinguished coursework. That is not to say that bad grades or a low GRE score should end the discussion on a given applicant , simply that a weak academic record should raise a red flag.
Also, research experience matured as an undergraduate student should be evaluated with care, and a grain of salt. The scope of the project and/or the real involvement of the student are usually necessarily very limited, given the short time available for the completion of the work (one or two semesters during which the student takes a full load of courses). Having written a routine in C to interface a piece of equipment with a computer, solved a complicated equation of unclear origin using Mathematica, made a bunch of good looking plots, and so on and so forth, are all valuable things to do, but should not be taken too seriously as telltale signs of research excellence.
2) The contention that at the age of 22 one should know whether experimental particle physics or theoretical condensed matter is the place for him/her to be, otherwise (s)he is likely lacking focus or motivation, is just gratuitous. I myself, for example, started out as an experimentalist, then decided that I liked doing calculations better and switched to theory, and have seen countless other students do the same or the opposite. In most cases this was not a traumatic event for the department or for the student — maybe the initial supervisor was disappointed, but for the most part people prefer not to continue to supervise students whose heart and mind are elsewhere.
In fact, precisely because the relationship between a graduate student and the initially chosen supervisor often does not turn out as profitable as the two of them expected at the outset (for a number of possible reasons), it is to the advantage of the program to admit students with broad interests and a strong academic background, enabling them to look for another supervisor, possibly in a different area altogether.
A student who was admitted into the program with the “tag” of a specific academic or group, is often unable to switch and drops out altogether, if things do not work out as originally planned.
 It is often the case that this type of applicant has contacted that individual faculty before submitting a formal application to the program, inquiring whether said faculty would be interested in him/her as an advisee. If interested in supervising the applicant, in turn, the faculty will typically requests the GAC that such application be ‘earmarked’, i.e., that special consideration be given to the applicant in the selection process.
 Too late. I can already hear keyboards clicking of commenters who are eager to attribute to me that very statement ;-)