Many school reform initiatives now require school leaders to make curriculum and resourcing decisions on the basis of evidence about their likely impact on student achievement. In this paper, we discuss the implications of this policy demand in the context of a descriptive study of schools' capacity to collect, analyze, and use collective achievement data to inform curriculum review and revision. The study of 47 teacher leaders in 26 elementary and middle schools showed that there was extensive assessment of collective achievement in literacy. While most schools' data were adequate to evaluate regular programs in reading, they were less adequate for the evaluation of writing, spelling, and extra reading programs. School leaders' evaluations of program success were not strongly linked to student achievement, and when they were, their judgments were frequently not data-based. When data were used, it was likely to be incorrectly interpreted. The implications of the study are discussed in terms of how reform policies and initiatives can build local capacity for data-based review and revision of teaching programs. (Contains 6 tables and 1 figure.)