Richard Halverson

Abstract Recent policy demands for external accountability are challenging instructional leaders to rethink how they have traditionally guided the practices of teaching and learning in schools. I will discuss how a new conception of instructional leadership is emerging across the country that focuses on how local leaders and teachers build data-driven instructional systems (DDIS) in their schools. I will argue that a learning sciences perspective, grounded in distributed cognition and the situational distribution of leadership practice, provides a unique perspective for accessing and representing how school leaders are building socio-technical systems to facilitate information flows about student achievement in their schools. The DDIS model describes how leaders create and use artifacts to coordinate a series of organizational functions involving 1) data acquisition, 2) data reflection, 3) program alignment and integration, 4) instructional design, 5) formative feedback and 6) test preparation. To illustrate how DDISs work in practice, I will review data collected in a year-long study of four schools to describe how the leaders structure opportunities to engage in data-driven decision making. Rather than reducing teaching and learning to mechanical processes as feared by many critics of accountability policies, I will show how the collaborative design work involved in DDIS assembly results in powerful professional communities well-suited to address chronic problems of instructional practice in schools.