Some of the most important evaluation questions cannot be answered through experimental research. The "assessment" of the St. Louis Leadership Program (SLLP), sponsored by the Danforth Foundations, is used as an illustration of a "connoisseur-based study" approach to evaluation. The project team, given limited time and funds, decided that it could not determine the effectiveness of the SLLP per se. However, it could and did do the following tasks in assessing the quality of SLLP: (1) ask participants for their opinions about the program, interview Danforth staff, examine program files, and observe aspects of the 1979-80 program in operation; (2) describe the selection and other program processes from an external perspective; (3) analyze interrelationships among program components; (4) interpret the findings; and (5) make recommendations based on interpretation of the findings. Through observation, description, and judgment, some defensible and useful advise was given, but was it research? Judgment, rather than some of the methods and standard approaches, must be given a central place in evaluation. (RL)