The purpose of this article is to suggest ways of using research on children's reasoning and learning to elaborate on existing national standards and to improve large-scale and classroom assessments. The authors suggest that "learning progressions"--descriptions of successively more sophisticated ways of reasoning within a content domain based on research syntheses and conceptual analyses--can be useful tools for using research on children's learning to improve assessments. Such learning progressions should be organized around central concepts and principles of a discipline (i.e., its "big ideas") and show how those big ideas are elaborated, interrelated, and transformed with instruction. They should also specify how those big ideas are enacted in specific practices that allow students to use them in meaningful ways, enactments the authors describe as "learning performances". Learning progressions thus can provide a basis for ongoing dialogue between science learning researchers and measurement specialists, leading to the development of assessments that use both standards documents and science learning research as resources and that will give teachers, curriculum developers, and policymakers more insight into students' scientific reasoning. The authors illustrate their argument by developing a learning progression for an important scientific topic--matter and atomic-molecular theory--and using it to generate sample learning performances and assessment items. (Contains 1 figure and 13 footnotes.)