Two aspects of educational research--that children's learning is typically the object of inquiry and that this learning usually occurs in organizational settings of classrooms and schools--locate educational research at the nexus of two of the most troublesome and persistent methodological problems in the social sciences: the measurement of change and the "unit-of-analysis" problems. Over the past several years, a satisfactory solution to these problems has been developed through the use of hierarchical linear models. This article summarizes these developments and shows how these techniques have already been used to illuminate important substantive questions about the effects of school organization and children's language development. In the final section of the article, the authors introduce a three-level hierarchical model that they argue should constitute the basic paradigm for future quantitative research on student learning.