This study tested the validity of the prevalent assumption that a teacher's understanding of the nature of science, in and of itself, is related to changes in the conceptions of his/her students and identified the classroom dimensions (such as teacher behavior and classroom ecology) significantly related to changes in students' conceptions. Data were obtained from pre- and post-test administrations of the "Nature of Scientific Knowledge Scale" (NSKS) to both teachers and students. Three intensive qualitative field observations in each of 18 biology classrooms yielded 44 variables of teacher behavior and classroom climate. The data collected from the observations were compared with the degree of change (high or low) in teacher and student views as measured on the NSKS. Results indicate that teachers' conceptions of the nature of science are not significantly related to changes in students' conceptions of science. Results also indicate the existence of several common characteristics in classrooms in which students exhibited the largest conceptual changes. For example, problem-solving and frequent, inquiry-oriented questioning typified these classrooms. In addition, students in these classrooms were more attentive and exhibited more participation. (JN)