Graduate Education: Deep? Broad? Both? Neither?

A perennial controversy, off and on, within every graduate program in psychology, concerns requirements for the PhD. What should they be?

Prior to answering this question, we need to ask: “Do we know the best way to train the next generation of researchers and scholars?” I think the answer to this question is doubtless “no” – hence our perpetual discussion of the issue with no resolution. I suspect the basic arguments (and techniques) for graduate training haven’t changed much in the past century, so this column could have been written in 1903. The discussion seems to go on within departments without often being aired in national publications, so this column will start the conversation in the Observer. Feel free to weigh in with letters to the editor at [email protected].

On the one hand is the strict mentor system of graduate education. At its most extreme, a young psychological scientist apprentices her- or himself to a senior scientist and learns about performing research (and anything else worth knowing) from this one person. The senior scientist will have a lab group, so others might be involved in the student’s education, but in some cases the strict mentor system amounts to a kind of home schooling at the graduate level. The student learns what the mentor and the lab group know but isn’t much permitted to learn from others. This pure form of the mentor system is, I think, more rarely seen in psychology than in other sciences, but it certainly exists. I have painted a caricature here, but actually this system often works to produce outstanding, quite prolific scientists. Its products leave graduate school with many publications (there is more time for research if the student doesn’t take many courses), and the students therefore get good jobs, because the number of publications is often correlated with success in the job market. The fact that these psychologists are narrowly specialized and know little about many aspects of the field (even ones near their own area) does not hurt them, at least in the short term. I know some excellent psychologists who were trained like this and make no bones about the fact that they know little about most of psychology (and that may be more than they want to know).

At the other extreme is the graduate program that seeks broad training and requires many courses to cover the breadth of psychology. Not only are students required to take statistics, research methods, and courses in their specialty area within the field, but they are also required to take many other courses for breadth, to insure they know about the field as a whole. They have a mentor for their research and for general advice, but the program is probably less research-intensive than the pure mentor program. Much of the education comes in classroom settings. These broad programs often also have a rather fearsome qualifying examination in which students are expected to know about many fields of psychology in some depth. The danger in such programs is that students might be broadly trained in courses but not have enough time to become expert in research in one area in the field. They may leave graduate school qualified to be good teachers, but with no or few publications, and without intensive research training.

Most graduate programs try to steer between the Scylla of narrow mentoring and Charybdis of a vacuous general program. These hybrid programs seek to have some distribution course requirements (but not too many), so students can be guaranteed to know something about various aspects of psychology. However, the program still has strong mentoring and provides adequate time to foster development of a good program of research. To attain the best jobs, contemporary students must gain large numbers of publications relative to years past. (A friend of mine commented that the level of productivity that once earned tenure is now required to get a job.) The graduate program at my university follows this hybrid model, and I think the program is a good one, but we still have discussions about the particular courses required (e.g., we require history of psychology or history of neuroscience, and not everyone is a fan of this idea), the flexibility of the requirements (should distribution courses be tailored to a student’s interests and program?), and so forth. The debate continues on many fronts, and we are examining our requirements again this year.

I have listened to discussions about graduate requirements over the years at several universities and now have formulated a fundamental law: Whatever graduate requirements faculty had as graduate students are the ones they endorse for use in the department’s requirements (a kind of intellectual imprinting). If someone endured a set of graduate requirements that bordered on intellectual hazing, then this person usually believes that these same procedures should be inflicted on new generations of students. If one had something different, then that system is preferred.

Perhaps it is for this reason that I have always thought that graduate requirements should be light. I attended Yale in the late 1960s and early 1970s, the Age of Aquarius, and at that time Yale’s graduate requirements were few. As I recall it, we had to take a one-year statistics sequence and maybe a few other courses of our choosing, and that was about it. As a qualifying exam we did have to read three books and write essays on them, a requirement reminiscent to me at the time of my sixth grade book reports.

At least, I think these were our requirements at Yale. I am recollecting them with a 30-year retention interval, so I may be wrong. (After all, memory is a tricky business – see Roediger & McDermott, “Tricks of Memory,” in Current Directions in Psychological Science, 2000). I do recall taking lots of courses at Yale in social and cognitive psychology, and a great course from Allan Wagner in animal learning, too, about the time the Rescorla-Wagner model was brand new. I took courses because I figured it was my only chance to learn from the outstanding faculty there. However, I followed this strategy by choice, while other students chose to take hardly any courses at all. I thought then (and think now) that I got a great graduate education at Yale, both in terms of coursework and in terms of research mentoring (with Robert Crowder and Endel Tulving). The students at Yale designed their own educations, with some advice from their mentors and other faculty. Isn’t this how it should be? Or would this system fail at other places?

To return to my original theme, I suspect that one reason that there is no universal agreement on the proper way to educate the next generation of graduate students is that we seek a “main effect” solution when the answer is a complex interaction: Some faculty are excellent mentors in one system but not others, and some students will greatly profit from some systems of graduate education but not others. The trick is finding the match between programs, mentors, and students, and this process can probably only be discovered by trial and error. So, a column like this one, debating these issues, will probably appear in the Observer in 2103.