Research has established that students enter their science classes with ideas about the natural world that do not align with accepted scientific beliefs. The diagnosis of these student preconceptions may be seen as an initial, crucial step in the process of teacher-facilitated conceptual change. So as to capture what science teachers do in their everyday classroom routines to diagnose these student preconceptions, in-depth observations and interviews were conducted with four exemplary secondary science teachers. The teachers' strategies for diagnosing students' preconceptions, their use of information found through diagnosis, and the teachers' understanding of students' preinstructional ideas were all analyzed. The four teachers in this study did not use any formal assessment tools such as pretests, concept maps, interviews, or journal writing to diagnose students' ideas, although they all stated that finding out what students know prior to instruction is important. The teachers all declared that they used questioning to gather information on student ideas but were seen in their classes to rely heavily on low-level, recall questions. One of the teachers, the most experienced, did conduct class discussions where the students were encouraged to express their ideas and preconceptions. The four teachers' understanding of what student preconceptions consist of and the justification for attempting to diagnose these ideas was seen to be weak. The implications of these findings and recommendations for teacher training are made in the article.