Richard Elmore was one of the most consequential scholars of educational leadership and school improvement of his generation, whose career combined serious academic inquiry with deep, sustained engagement in the practical work of schools. He served as the Gregory R. Anrig Professor of Educational Leadership at Harvard Graduate School of Education from 1990 until his death on 9 February 2021, and in that role he shaped the thinking of hundreds of aspiring educational leaders who passed through the doctoral programme he eventually directed, the Ed.L.D. He had previously taught at Michigan State University and the University of Washington, and he served in the federal government as a legislative liaison on education policy issues, an experience that left him with a keen sense of the gap between the intentions of policymakers and the realities that those policies encountered in classrooms. His early research grew from two persistent questions: why do the actual consequences of educational policies bear only a vague resemblance to the ideas that shaped their design, and why do federal programmes designed to test the effectiveness of different instructional approaches find more variability within a given approach than between approaches? These questions oriented his life's work toward the instructional core of schooling — the relationship between teacher, student, and content — and away from the macro-level policy prescriptions that dominated educational reform discourse. Elmore spent at least one day per week in schools throughout his career, working directly with teachers and administrators on the practical challenges of instructional improvement; colleagues recalled that he had a sharp intellect matched by a generous heart, a contagious laugh, and a profound respect for the capacity of young people. He was a member of the National Academy of Education and a past president of the Association for Public Policy and Management.

Elmore's most fundamental argument was that meaningful educational reform must begin “from the inside out” — from the knowledge, skill, and professional culture of teachers and school leaders, not from external mandates, performance standards, or accountability systems imposed from above. He observed that decades of school reform in the United States, from the accountability movement of the 1970s through No Child Left Behind in 2001 and Every Student Succeeds Act, had consistently failed to produce sustained improvement in the quality of instruction experienced by students, particularly students of colour and students living in poverty. He traced this failure to what he called the structural problem of American schooling: that teachers work in isolation behind closed classroom doors, that the quality of instruction is treated as a private matter of individual style and artistry rather than a shared professional responsibility, and that the organisational structures of schools actively prevent the development of a common culture of practice. Reform imposed from outside, he argued, pushes against this structure without changing it, generating compliance behaviour rather than genuine improvement.

Elmore's concept of the Instructional Core — the relationship among teacher, student, and content — provided educational leaders and reformers with a deceptively simple but analytically powerful framework. His argument was that everything that happens in a school system is relevant only insofar as it affects what happens in this triangular relationship: if a policy, programme, or initiative does not change what teachers do, what students are asked to do, or the level of complexity and challenge at which they engage with content, then it will not improve learning outcomes, regardless of how well intentioned or administratively efficient it is. He identified three levers for improving student learning within the Instructional Core: increasing the teacher's instructional knowledge and skill, increasing the level of complexity of the content students are expected to master, and changing the role of the student in the learning process, shifting from passive reception toward active engagement with challenging material. This framework gave school leaders a way of evaluating proposed reforms not by their rhetorical appeal but by their direct impact on the core.

Drawing on the medical rounds model familiar from teaching hospitals, where groups of physicians and medical students observe patients together, discuss evidence, and develop shared diagnostic and treatment protocols, Elmore and his colleagues developed the process of Instructional Rounds as a mechanism for opening classroom practice to collegial observation and collaborative analysis. In an Instructional Rounds process, a network of educators — teachers, principals, central office administrators — visits a series of classrooms in a school facing a specified “problem of practice,” collects observational evidence without making evaluative judgments, and then engages in structured analysis to develop a shared understanding of the problem and generate potential solutions. The process was explicitly designed to break down the norm of privacy that Elmore identified as the central structural obstacle to improvement: by making instruction visible and subjecting it to collegial scrutiny in a non-evaluative context, Instructional Rounds build the shared language, mutual trust, and professional culture of continuous improvement that external accountability systems cannot produce. Elmore described the process as developmental and inherently messy, requiring long-term commitment and a willingness to tolerate the discomfort of having one's practice examined.

Through his direction of Harvard University's Public Education Leadership Project (PELP), Elmore and his colleagues developed the Coherence Framework as a visual representation of the complex relationships among the elements that affect the Instructional Core: the culture of the school and district, the organisational systems and structures through which decisions are made and resources allocated, the human capital strategies through which teachers and leaders are recruited, developed, and evaluated, the stakeholder relationships through which schools engage with families and communities, and the broader external environment of policy and accountability. The Coherence Framework's central insight is that improvement strategies fail not because they are technically flawed but because they are pursued incoherently — because different elements of the organisation pull in different directions, because the culture of the organisation resists the new practices the strategy demands, or because the leaders who design the strategy fail to consider its effects on the people who must implement it in classrooms. Backward design — beginning with the desired change in classroom practice and working outward to determine what organisational conditions must change to make that practice possible — was Elmore's recommended antidote to the incoherence that undermined so many reform efforts.

Elmore's Leaders for Learning MOOC, delivered through HarvardX, represented his effort to bring his framework for educational improvement to a global audience of practitioners, policymakers, and citizens interested in the future of learning. The course organised perspectives on learning along two axes — hierarchical to distributed, and individual to collective — generating four modes of learning: Hierarchical Individual (the traditional transmission model of schooling), Distributed Individual (self-directed digital learning), Hierarchical Collective (social-cognitive development in group settings), and Distributed Collective (self-organised learning networks based on expertise and shared interest). Elmore used these modes not to advocate for one over the others but to prompt participants to examine their own assumptions about what learning is, how it occurs, and what forms of organisation best support it. He argued that most educational reform failed because reformers were operating from one theory of learning — typically the Hierarchical Individual model — while the students they sought to serve were increasingly operating from the Distributed Collective model, creating a structural misalignment that no amount of additional accountability pressure could resolve.

• Restructuring Schools: The Next Generation of Educational Reform (1991)
• School Reform from the Inside Out: Policy, Practice, and Performance (2004)
• Instructional Rounds in Education: A Network Approach to Improving Teaching and Learning (2009, with Elizabeth City, Sarah Fiarman, and Lee Teitel)
• I Used to Think…and Now I Think: Twenty Leading Educators Reflect on the Work of School Reform (2011, editor)
• Learning by Design: Live, Play, Engage, Create (2019, with Prakash Nair and Romi Doctori)