In the early 1990's, the University of Indiana computer science department put together an excellent series on "Survival in the Academy." Most of the material is still very much up to date and much of it is applicable to graduate students and faculty in all disciplines. Below is a section on whom to look for and whom to avoid in a graduate research advisor:

Whom to look for

• Someone with similar interests - Seek someone with whom you share research interests; otherwise, you may undertake a project that you do not value and never complete it.
• Someone with compatible interests - If you cannot work with the research group whose research goals are most similar to your own, you may nonetheless learn a great deal working with another group. It is quite possible, of course, that as you work in the laboratory of "second choice" you may become interested in the research problems there.
• Renowned researchers - Seek people who love science and their research. They will document their work in articles, published in respected journals, that often describe a series of inter-locking experiments concerned with a single problem. When researchers value their work and others agree, others will extend the work. Invited articles and presentations to professional societies suggest that a researcher's work is well-received.
  
  Grant support from major research foundations, such as the National Science Foundation, indicates that other scientists judge this person to have made significant contributions. Such grant support is allocated competitively, more competitively than is space in major journals. A history of grant support from major foundations is, therefore, very impressive. Most impressive is a researcher who holds a special position where a university or a foundation has granted the person a lifetime of research support.
  
  There are potential problems working with renowned researchers, however. In areas where research costs require a lot of grant support, such advisors may be limited in the amount of attention they can devote to you because they are busy writing grant proposals, justifying grants, administering grants, and supervising post-doctoral students.
• Beginning researchers - Younger or less famous faculty often make excellent advisors. In the absence of substantial grant funding or a lengthy publication list, look for someone who has an active and growing research program.
  
  It is possible to have the best of both research types. If there are renowned researchers in your department, you can include them on your research committee (with your advisor's consent), seek their advice and eventually, if all goes well, seek their letters of recommendation and "connections."
• Someone you can respect - Choosing or keeping an advisor primarily because he or she is nice is a mistake. Sometimes nice people withhold frank evaluations of your knowledge, skills, and progress to avoid hurting your feelings. If you have an excellent advisor, your feelings toward your advisor might best be labeled as respect.

Whom to Avoid

• Grant swingers and research millers - Do not equate grant support or the size of an institute or research laboratory with quality. Not all research is expensive. Don't eliminate an advisor based solely on his or her grant support. Seek an advisor who knows quite a bit about your area, is enthusiastic about research, has a good publication record, and, of course, seems willing to offer help.
• Those not at the bench - The best advisor is one who is still involved in active research themselves. The American Association of Universities recommends that "students must be directed by experienced scientists. The director should supervise, teach, and encourage in-depth scrutiny and interpretation of results, emphasizing respect for primary data. Routine audit and review of all primary data by the laboratory director is strongly recommended. It is inadvisable for the director to delegate these important functions" [Report of the Committee on the Integrity of Research, 1983].
• The perpetual adminitrator - If your advisor of choice is someone who has many other professional responsibilities (being an officer in a professional society, the departmental chair, or editor of more than one journal), verify to your own satisfaction that this faculty member will have adequate time to work with you. Professional service activities are important and contribute to others doing science, but they can substantially reduce supervision quality, unless you are only one of a few advisees. If you think this might be a potential problem with your prospective advisor, ask him or her about it.