Three mathematics teachers and 10 of their ninth grade students were observed and interviewed during a 6-week period. One teacher claimed that her main challenge--implementing a student-centered curriculum--was her doubt that students would make the right connections without her explanations. Another teacher struggled with the dynamics of operating both small-group and whole-class discussions and ultimately decided not to hold whole-class discussions. A third teacher achieved a more equal balance between teacher-directed and student-centered activities. All three teachers demonstrated more difficulty than students changing their expectations about appropriate mathematical activity. (Author)