Using Agent-Based Simulation to integrate micro/qualitative evidence, macro-quantitative data and network analysis
12 May 2014
BSA Social Network Analysis Group - Mixed Methods Approaches to Social Network Analysis Conference (organised with with University of Greenwich and Middlesex University)
Middlesex University, UK
Networks are an abstraction of complex social processes. Albeit themselves formal, the social processes on which they are based can be researched using both quantitative and qualitative methods. The problem in combining these approaches comes from the very different natures and levels on which they are based. Here we describe an approach which uses agent-based modelling (ABM) as a stepping stone towards the more abstract network models. These ABMs are more in the nature of complex and dynamic descriptions than general theories, and are ideally suited for integrating a variety of kinds of evidence into a coherent fashion - including quatitative evidence to inform the micro-level behaviours of agents, and quantitative evidence about the macro, aggregate levels. The assumptions behind these kinds of ABM are relatively transparent, and the ABMs used to generate networks in a precise manner. Thus this "staging" of the abstraction process allows a well-founded mixed-methods approach to social network research. A worked example of this on voting behaviour is presented.
Slides at: http://www.slideshare.net/BruceEdmonds/using-agentbased-simulation-to-integrate-microqualitative-evidence-macroquantitative-data-and-network-analysis