Many sociotechnical phenomena are dynamic, nonlinear, and difficult to explain with static models alone. We develop and use computationally intensive approaches—particularly agent-based simulation and combinations of simulation with machine learning and LLMs—to construct and elaborate theory.
These approaches allow us to formalize mechanisms, explore interactions across multiple levels, and examine how system-level outcomes emerge from local actions and feedback over time. Computational methods complement, rather than replace, qualitative and empirical inquiry in our theory-building process.