“Agent based modeling: A method for understanding individual, social, and environmental influences on learning and equality in the classroom.” Learning science in the past decade has provided considerable evidence that learning is at once emotional, social, and cognitive (MH Immordino-Yang, et al 2018). As we seek to improve social, emotional and learning outcomes in schools around the world, it is important to better understand how individuals adapt to and influence each other and their environments as they connect and interact daily in and outside the school and classroom. Furthermore, it is important to develop a better sense of how local interactions shape education. How do individual interactions shape the patterns of learning outcomes in a school or at a larger scale such as a district? How do they shape the nature of the learning environment and in turn, how do differential learning environments shape the patterns of interactions and relationships in a school? This information holds enormous potential to inform international education programming that may hold promise for improved uptake of innovations and education outcomes. How interactions locally shape education – in schools, administrative offices, or higher education institutions – is not well understood or studied extensively. Agent-based modeling (ABM) is a technique that can be applied to better understand the link between local dynamics of individuals in a school or community and certain aggregate characteristics of a school or emerging school changes. ABMs are based on the application of algorithms or simple rules representing the lower-level interactions of individuals (or system components) that give rise to higher level system structures or changes, providing a tool to understand bottom-up influences on education outcomes. (See M Macy & R Willer, 2002; M Jacobson, et al, 2017). In this presentation we present an agent-based model to show the potential impact of teacher feedback on student participation in the classroom and the relative impact of students who are more or less vulnerable (e.g., have lower/higher ability levels and are from more/less marginalized backgrounds). The model was informed by student data from primary schools in Uganda and Tanzania. The model demonstrates that over time, when met with repeated experiences of negative feedback, more and more students will quit participating entirely and some will dropout, especially children who are more vulnerable. On the other hand, when teachers are increasingly positive, more and more students participate more actively, even among the most vulnerable children. Thus, the nature of teachers’ responses to students when they answer questions in class can powerfully impact student participation and shape equality in participation. To extend, this would seem to impact student learning. The objectives of this presentation are: 1. To introduce the agent-based modeling method. 2. To present an application of ABM in international education 3. To demonstrate the utility of ABM in research, policy dialogue, and programming.