School districts across the country have adopted interim and benchmark assessments in response to NCLB pressures to raise student achievement, despite the lack of a research base. Thus, it is especially important that well-conceived, empirical studies of the effects of such programs be carried out. Various theoretical frames can help understand how teams of teachers might begin to use data to reflect upon and adjust their instructional practices. At issue is whether an idealized data-based decision-making theory of action will play out in practice. Or, will teaching-the-test practices instead be exacerbated? What we learn from the empirical studies reviewed here is that positive examples, where assessment results were coherently linked to curriculum and instruction, were facilitated by highly committed principals and teacher leaders. But, these examples were rare. More frequently, interim assessments results appeared to be used, item by item, to reteach steps in problems that were missed without attending to underlying concepts or gaining diagnostic insights. (Contains 1 figure.)