Background/Context: Students, classrooms, teachers, and schools form hierarchical units of any schooling system that compete for, say, educational resources and sometimes blame one another for the failure to meet a certain societal goal. Little educational research exists that separates competing effects on schooling outcomes across these educational levels. Purpose/Objective/Research Question/Focus of Study: We aimed to fill in this gap in the literature by simultaneously examining in one model students, classrooms, teachers, and schools for competing effects on science achievement and identifying factors at each of these educational levels critical to science achievement. Population/Participants/Subjects: We used data from the Program for Regional Assessment and Promotion of Basic Education Quality, a program in China. Data described learning and teaching practices of eighth-grade students and their science teachers (physics, biology, and geology) and school principals. Research Design: A four-level hierarchical linear model was developed to perform secondary analysis of science achievement data with students nested within classrooms, nested within teachers, nested within schools. Findings/Results: Across physics, biology, and geology, the proportion of variance in science achievement attributable to students was 75%-78%, 7%-9% to classrooms, 1%-4% to teachers, and 10%-14% to schools. Across the three content areas of science, male students, younger students, and students with higher parental socioeconomic status outperformed their counterparts respectively (at the student level), and students in schools with higher student expenditure outperformed students in other schools (at the school level). Teacher effects were diverse across the three content areas of science. Conclusions/Recommendations: The breakdown of institutional contributions to science achievement across the three content areas of science indicates a hierarchy of importance of educational units in the order of schools, classrooms, and teachers in China. From a global perspective, in China, gender differences were stronger, socioeconomic differences were weaker, age effects deviated from the international tradition, and school effects associated with student expenditure were stronger. We conclude that preparation and development are as critical for school administrators as they are for classroom teachers and that professional development organized by science content area may be more effective than pooling all science teachers together.