One assumption underlying accountability policies is that results from standardized tests and other sources will be used to make decisions about school and classroom practice. We explore this assumption using data from a longitudinal study of nine high schools nominated as leading practitioners of Continuous Improvement (CI) practices. We use the key beliefs underlying continuous improvement - derived from educational applications of Deming's TQM models - and organizational learning to analyze teachers' responses to district expectations that they would use data to assess their own, their colleagues', and their schools' effectiveness and to make improvements. The findings suggest that most teachers are willing, but they have significant concerns about the kind of information that is available and how it is used to judge their own and colleagues' performance. Our analysis reveals some cultural assumptions that are inconsistent with accountability policies and with theories of continuous improvement and organizational learning. We also identify barriers to use of testing and other data that help to account for the limited impacts.