This report summarizes results from research aimed at improving the quality of information collected about school curriculum. The research sought to design and pilot a model for collecting benchmark data on school coursework. These more indepth data, such as course textbooks, assignments, exams, and teacher logs, can serve as anchors against which the validity of the survey items used in national data collections might be assessed. The data provide a basis for assessing the extent to which survey items measure what is taught in schools and classrooms. They can also be used to monitor whether the validity of teachers' responses have been undermined by outside factors. Data were derived from a survey of 70 mathematics teachers in 9 secondary schools located in California and Washington. The survey was administered before and after the collection of artifact data. Data were also collected from teacher daily logs, assignments, and interviews with principals, counselors, and mathematics department chairs. Chapter 2 details the study design, and the next three chapters summarize the extent to which major dimensions of curriculum can be measured through national surveys and then validated through deeper probes in a smaller number of sites. The final chapter discusses the implications for the design of future curriculum-indicator systems and for the policy uses of such information. It concludes that while an enhanced version of current national surveys can provide a reasonably accurate picture of high school mathematics teaching across the country, there are significant limitations on such data, and at this point, policy uses for more than informational purposes would be inappropriate. The study represents a first step in ensuring that curriculum indicators are valid and reliable measures of instruction. Nine figures, 12 tables, copies of the surveys, and a sample daily log form are included. (Contains 39 references.) (LMI)