The problem-solving mode of teaching science implies change for teachers, administrators, state education departments, and other individuals charged with implementing educational innovation. This study provides a descriptive record and interpretation of the words and actions of an elementary school teacher beginning to learn about teaching science as problem solving rather than dispensing content information, and attempting to modify her teaching accordingly. Data collection spanned a period of one year and included classroom observations, interviews, and document analysis. Observations revealed that environmental factors such as administrative support and flexibility, availability of science materials, school philosophy, parental support, and teacher status strongly affected the teacher's effort to change. Personal or internal factors also influenced the teacher's classroom practice: background in science; ability to see interdisciplinary teaching possibilities; organizational ability; regard for individual student's ideas; need to maintain control over student activities and thinking; regard for other teachers' intelligence and experience; emphasis on success; need to "cover" a textbook; view of the relationship between science content and problem solving; and general openness to change. This study demonstrates that educators cannot control results merely by altering input. (MVL)