Developing a Learning Progression with Empirical SupportThe goal of this article is to present a proposed learning progression, organized around two dimensions,each ofwhich integrates practice and metamodeling knowledge.We refer to these key integrative dimensions as ‘‘Models as generative tools for predicting and explaining’’ and ‘‘Models change as our understanding improves.’’ We present illustrations of student performance and understanding along each of these dimensions drawn from several empirical investigations involving elementary and middle school students. We use excerpts from these data, which include written assessments of student reasoning with models, reflective interviews about modeling practice, and classroom discourse during modeling activities, to identify aspects of performance and understanding that are evidenced in students’ work, and those which may not be achieved. The data help elucidate the meaning of the dimensions of the progression, and help distinguish between various levels of performance. The data are not intended as a test of a particular curriculum or teaching approach, or an overall validation of the progression. Rather, the data help illustrate what kinds of performances and understandings in modeling are possible with elementary and middle grades students, with appropriate support, and how they can be characterized using the learning progression.