What’s the optimal composition of a graduate supervisory committee? I’m not sure. But here are some thoughts, please share yours.
It’s essential that all your committee members be decent, reasonable people who get along with one another and with you. You don’t want to get caught in the crossfire of fights between your committee members. And you don’t want your progress to be made unnecessarily difficult because one of your committee members has unreasonable expectations.
Beyond that, I think the ideal committee has a diversity of expertise and interests, and includes at least one smart generalist who doesn’t necessarily do anything directly relevant to the thesis. Anecdotally, I think students tend to overrate the importance of having committee members with specialized expertise directly relevant to their thesis research, and underrate the value of smart generalists. Your committee’s most important role is not to help you deal with purely technical issues like choosing the right error distribution for your GLM. Your committee’s most important role is to ask you good questions and offer good suggestions that you and your supervisor wouldn’t have thought of on your own and to which you can’t just look up the answer. Your committee members also are a good source of mentoring and career guidance.
For instance, my committee included Bob Holt.* I wanted Bob because he’d written the theoretical model I was planning to test for my thesis. This was a very understandable reason for wanting Bob on my committee. But in retrospect, it wasn’t a good reason. I didn’t need Bob to explain his model to me. But having a smart generalist like Bob suggest that I read Deborah Mayo’s book on philosophy of statistics just because it would be interesting and thought-provoking was hugely valuable to me. Having a smart generalist like Bob write reference letters for me was hugely valuable when it came time to apply for faculty positions. Having a smart generalist like Bob become a friend and mentor remains hugely valuable to me today. Unfortunately, “put Bob Holt on your committee” is not scalable advice. But “put a smart generalist like Bob Holt on your committee” is at least somewhat scalable advice.
Yes, it’s possible for a committee to be too broad, so that nobody really understands what you’re doing well enough to give you useful advice. And yes, there are circumstances in which it’s essential that a committee include particular technical expertise. For instance, several of my students have worked in systems in which I’ve never worked myself. I always make sure those students have someone on their committee who has expertise in the chosen system, since otherwise there’d be a serious risk of technical mistakes. But typically, I don’t think it’s the best idea to think of your committee as a bunch of technical consultants. You and your supervisor ordinarily will be the technical experts on the committee.
*He was my external committee member. This gets into another topic that’s probably worth a separate post, because the role of the external committee member(s) varies from place to place. Here at Calgary, the external examiner only joins the committee a few weeks in advance of the defense and their only role is to read and evaluate the thesis. I think that’s how things are usually done in Canada (Canadian colleagues: please correct me if I’m wrong on this). I hate this way of doing things and much prefer the US system. I think the Canadian system cheats students out of an important source of advice and mentorship without any compensating benefit. I presume that the Canadian practice is meant to ensure the objectivity of external examiners. But in practice I don’t think US external committee members are any less objective during thesis defenses than Canadian external examiners are.