Purpose: The purpose of this paper is to provide an empirical comparison of two measures of school success -- a value-added assessment system and the federally-mandated system of adequate yearly progress (AYP) -- to identify highly effective urban schools in the USA and to explore the predictive relationship between evidence-based decision-making and school improvement. Design/Methodology/Approach: A total of 204 urban schools with 6,684 teachers and 149,665 students in grades 1 through 10 participated in the study. Data included teacher survey and students' standardized reading and math scores from 2002 through 2005. Analyses included factor analysis, growth modeling, and multiple logistic regression analyses. Findings: AYP status was strongly predicted by student and school demographics rather than by organizational climate and instructional practices. In contrast, school growth as measured by the district's value-added assessment system was unrelated to the demographics of the student population and related strongly to specific school practices. Specifically, high growth schools exhibited strong evidence-based decision-making practice where teachers used the district's benchmark assessment to reflect on instructional practice, used the core curriculum to guide instruction, and received frequent and high quality professional development on reading and math instruction. Practical Implications: As states gravitate away from relying on AYP status as a measure of school success, districts will benefit from integrating measures of growth and using school data management systems that integrate benchmark assessment capabilities and provide teachers with the training and tools needed to use the information in their daily practice. Originality/Value: This study provides a direct comparison of evaluation models using a variety of current methods within a single district that has played a central and highly-visible role in the education reform movement in the USA. (Contains 8 tables.)