In a randomized controlled trial of a teacher data-use intervention, the Linking Study tested the impacts of a cyclical and collaborative process that linked teachers' data on instructional practice with data on their students' learning. This article describes the theory of the intervention and its roots in the literature as a backdrop for an intervention with 64 teachers in 27 professional learning communities and their students. The researchers found moderate and significant effects on external judgments of the quality of instruction caused by the intervention and small and marginally significant impacts on student performance on end-of-unit tests. There were no effects on teachers' perceptions of the importance of data or their selfreported proficiency using data. The article concludes with a discussion of the ways that data-use interventions might be strengthened to support instructional improvement and student learning.