Computational approaches in rigorous sociology: agent-based computational modeling and computational social science

Abstract

We discuss two computational approaches of particular significance for rigorous sociology, Agent-Based Computational Modeling (ABCM) and Computational Social Science (CSS). CSS exploits novel sources of large-scale data from, for example, the Internet, telephone-communication records, or population register data, studying digital traces of social interaction with great precision. ABCM uses computer simulation to study how social regularities can arise from complex interactions among interdependent individuals. CSS and ABCM share a strong focus on computational methods, but only a few contributions combine them. We briefly introduce both approaches and argue that they can strongly benefit from more interaction and integration. ABCM can address the generalizability and theoretical explanation of empirical patterns identified in CSS, whilst empirical CSS can help validate and calibrate otherwise abstract models from ABCM. We discuss examples of a successful combination of both approaches and address directions for more theoretical and methodological integration in the future.