2008-2011 | Merging Agent-based Modeling Techniques and Ethnography: A New Analytic Tool for Studying Illicit Drug Use Behaviors, Markets and Economies

The National Science Foundation, Cultural Anthropology & the Methods, Measurement, and Statistics Program (BCS-0724320)

Principal Investigator: Lee D. Hoffer

Research Team: Burchan Bayazit, David Epstein


This project will simulate the operation of the local heroin market in Cleveland, OH by synthesizing ethnographic decision tree and Agent-Based Modeling (ABM) techniques as an innovative strategy for improving ethnographic analysis and utilization. The specific aims of this research are to: 1) conduct an ethnographic study of the local heroin market that facilitates ABM development, 2) enrich these data using Ecological Momentary Assessment (EMA) technologies, and 3) construct a social / economic simulation computer lab that reproduces how the market operates. This proposal extends my on-going feasibility study, using ethnographic data to retrospectively construct ABM simulations of the local heroin market in Denver, CO., circa 1998-2000, and represents an essential next step in the evolution of this hybrid methodology.

Recruiting active heroin users from the community, ethnographic research will delineate roles, as well as attitudes, beliefs and behaviors concerning decisions associated with the use, acquisition, sales, and distribution of heroin within the market. Formalized decision tree diagrams will elaborate reasons, constraints, and environmental influences associated with these activities. A second phase of research will then collect interactional and transactional data, validate decision tree models, and determine the distribution of decisions utilizing EMA. In conjunction with the ethnography, this self-report data collected daily from participants using a Palm Pilot™, will provide the robust data necessary for constructing ABMs.