The shift to a focus on implementation, and the science of schooling and classroom processes inherent in that focus, is a fundamental change and one with considerable promise. Turning researchers' attention toward the dynamics of activity in educational settings, this shift will require a deeper conceptualization of the influences on learning and development, one that is far more connected to understanding and modeling the dynamics of change and much less wedded to evaluating average program impact. Such a shift will require methodologies that enable modeling the activity of complex systems over time and increasingly will have to accommodate disciplinary expertise and funding that span sociology and neuroscience. Fundamen tally, we need to know much more about how schooling works and does not work as we simultaneously and systematically work to improve it.