This paper evaluates the procedures and results of the National Assessment of Educational Progress (NAEP) in science. A brief description of the 1981-82 assessment, funded by the National Science Foundation, precedes a discussion of strengths and weaknesses of the current science assessment program, conducted by Educational Testing Service. Major strengths include: high technical quality; accurate portrayals of the national picture; consistent procedures enabling comparison across time; appropriate content; policy impact; and identification of report group differences. Criticisms include: (1) the science assessment cycle has been unnecessarily irregular; (2) assessments tend to focus on traditional subject matter content and to collect little additional information; and (3) subject matter experts who possess a commitment to their field do not seem to have sufficient interest in the data to attempt the difficult task of analyzing unfamiliar data tapes. The assessment program could be strengthened by requiring contractors to include the services of experienced subject matter researchers into the budget proposals. Data should be regularly gathered, and the measurement of attitudes toward science should be improved. Lines of communication between NAEP and the community of educators is essential for continued support of the assessment and for maximizing the educational impact of the information being gathered. (LMO)