This paper describes the introduction of a problem-based learning program in a new doctoral program at the University of Texas--Pan American in 1998. Students were assigned the problem of studying the implementation of a two-way bilingual education in their Educational Leadership Laboratory and making recommendations regarding the use of one or more of 10 interventions. The paper describes this attempt at problem-based learning which was unsuccessful due to several problems: complaints by students, and underestimation of the importance of having students actively involved in and informed about the emerging structure of the program design. The author concludes that the problem-based learning approach should not have been assigned so soon and on such a scale. (Contains 15 references.) (DFR)