This project investigates how educational knowledge, resources, and practices spread within and across online and in-person communities to improve teaching and learning in STEM (science, technology, engineering, and mathematics). While evidence-based practices and tools often exist, they can fail to reach or influence educators and decision-makers. This project focuses on brokers, such as organizations and individuals that help move knowledge into broader use. By understanding how these brokers operate in both real-world and online social networks, the project seeks to improve how innovations are shared, adapted, and sustained across the U.S. education system. Ultimately, this research aims to support the design of more effective strategies to connect educational research with everyday practice, benefiting students, educators, and policymakers alike. To achieve these goals, the interdisciplinary team will use a mixed-methods research approach, combining social network analysis, machine learning, artificial intelligence, surveys, and interviews. It will examine how brokers navigate and shape multi-level networks—both micro-level (individual relationships) and meso-level (online networks)—to translate and spread evidence-based STEM education knowledge. The research will explore how these networks evolve, what kinds of structures make diffusion more successful, and which broker activities are most effective. Through analytical models, simulations, and field-based studies, th