Agents, networks, and markets: Three essays from a complex systems perspective on educational change

In the three essays that comprise this thesis, I use the ideas and tools of complexity science to investigate a particular case of why well-intended management and public policies often provoke resistance by the very system they are trying to improve: the case of educational change.;The first essay attacks directly the question of why it is so difficult for reform programs to take hold inside a school. Through a combination of simulation and fieldwork at a large high school engaged in a major change effort, this paper develops an agent-based model of organizational inertia and change that reveals how stasis can emerge even in an organization comprised of well-informed agents striving for improvement. A central finding is that incremental improvement at the micro-level of an organization can eliminate the work activities needed to support the early steps towards more radical, macro-level change---a dynamic simultaneously moderated by an organization's social network structure, the strength of social influence within subgroups, and environmental turbulence.;The second essay examines an idea that underlies a large variety of education reform efforts---the idea that social interdependence plays a crucial role in improving schools. I use social network data collected at large public high school to examine the role of a students social structure in determining academic performance, and find that the association between academic performance and the characteristics of a students peers is moderated by the density of ties among a students social relations; a finding that points to the importance of considering the network structure of structure of students relationships when diagnosing the impact of building stronger community in schools.;The third essay addresses the difficulties involved in evaluating the impact of choice-based reforms in education by developing a series of agent-based models that capture the dynamic processes involved in moving from a non-market to choice-based system. Using historical data from Chicago Public Schools to estimate and set initial conditions, analysis of the models reveals how differences in capacity constraints across districts can result in uneven treatment effects across school choice programs, and suggests an approach of indexing results across contexts. Simulations also reveal how an increasing focus on school quality at the household-level can eventually lead to a less competitive structure and lower levels of district achievement---a paradoxical mismatch between micro- and macro-levels of behavior.;Taken together, these essays contribute to our understanding of the dynamic cause and effect relationships that often make schools and school districts resistant to change, and provides insight into the application of a novel approach to studying social and organizational dynamics.