I was asked the other day what factors to consider when picking a thesis committee. And I realized that this is not a question I have pondered a lot. Normally, when one of my students needs to pick a committee, I just recommend people that seem a good choice. I don't have a properly thought out, systematic framework to steer the selection process. So with this post I’ll try to develop a more systematic approach to this question.
Let’s first think about the purpose of the thesis committee. There are at least four distinct functions the committee should perform: 1. The committee should be your personal science advisory board. It should provide relevant expertise you or your adviser may lack. 2. The committee should review your research approach and verify that your data support your conclusions. 3. The committee should make sure you progress adequately and stay on track. 4. The committee should serve as advocates on your behalf in case your adviser develops unreasonable expectations.
The advisory-board function is probably the most straightforward to satisfy. Pick a group of people that, jointly, cover the topics relevant to your work. For example, if you work on the evolution of influenza virus, you might want to pick an expert in the molecular biology of influenza, an evolutionary biologist, an epidemiologist, and a computational biologist or biostatistician . Of course, to some extent the choice may depend on the expertise of the lab you’re in. If your lab is an influenza lab then your adviser may be the influenza expert and it might be more important to have another evolutionary biologist. By contrast, if your lab is an evolution lab, it might be more important to bring in more influenza expertise.
For the remaining three points, personality of your committee members matters more than expertise. You want people who will speak up, who won’t let you get away with BS, but also who care for you and won’t put you through any unnecessary difficulties. Basically, people who can give you tough love. I wouldn’t worry too much about picking people who have a reputation of speaking their mind. Once you’re past candidacy, it’s unlikely that you will be kicked out of graduate school, and in any case if there are issues with your performance or research approach you’d want to hear about them earlier rather than later. In my mind, the worst committees are those that let a student bumble along for six years and then say “Well, this work doesn’t quite rise to the level that we expected from a PhD.” I think the best committee members are those that push you to develop a clear plan on how to complete your work, that tell you exactly what they expect from you before you can graduate, and that reign you in when your plans get overly ambitious (which happens to almost every student).
Next, consider possible inter-faculty dynamics. Don’t put two professors on your committee that are known to have issues with each other. You might end up as collateral damage in a fight between them . Also, make sure your committee has at least some members who could speak up against your adviser if necessary. If your adviser is a very senior scientist, choosing four assistant professors as committee members would be inadvisable. Have at least one, and better two or more, committee members of comparable rank and seniority.
Committee members to avoid are those that are overly passive, that like to talk just to hear themselves speak, and that tend to get lost in tangents or irrelevant minute details. Your adviser and your fellow graduate students should know who they are. Finally, it’s important that your committee members are actually available to you. The best committee doesn’t do you any good if you can never get them all into the same room at the same time. Therefore, I’d advise against any committee members who have a reputation for being difficult to schedule. It is generally known in the department who is never around, or who may rarely have more than one open time slot every few weeks. All else being equal, I would recommend against putting such a person on your committee.
 Every committee should have a computational biologist or biostatistician, to verify data is analyzed properly and results are statistically sound.
 I have never personally witnessed something of this sort happening, but I’m sure some poor graduate student somewhere is finding himself or herself in exactly this situation right now.