A Framework for Computational Ethnography

• In Event: The Promises and Pitfalls of New Technologies for Qualitative Research

Abstract

Agent-based modeling (ABM) has long been used in sociology and the social sciences to explore the macro-level consequences of abstract rules. ABMs typically begin with stylized assumptions derived from theory and examine the emergent patterns they produce. While powerful, this tradition has remained largely theoretical: the rules governing agents are rarely grounded systematically in empirical fieldwork.
This paper argues that there is substantial untapped value in relocating agent-based experimentation into the empirical realm through what I call computational ethnography. Rather than using ABMs to test abstract heuristics, computational ethnography begins with case-based fieldwork and treats observed mechanisms—interactions, relational dynamics, power asymmetries, and strategic practices—as the basis for model construction. Ethnographic data are conceptualized as structured relations and functions (“morphisms”) that can be translated into rule-based simulations.
By encoding empirically observed mechanisms into agent-based models, researchers can experiment on cases themselves: testing the stability of observed arrangements, exploring counterfactual conditions, and probing the scope and limits of emergent explanations. This approach enables systematic variation of contextual factors while preserving allegiance to the complexity of the fieldsite.
The framework rests on three principles—composition, modularity, and dynamics—which allow theories to be formally integrated, sub-mechanisms to be stress-tested, and feedback processes to be explored over time. Computational ethnography thus transforms the case from an illustrative narrative into a dynamic system for experimentation and ultimately generalization.

Author

• Vivek Ramakrishnan, University of California-Los Angeles