Achievement

Social networks are fundamental to social life, but network data for general populations are rare. One approach is to use data collected in social surveys to build indirect associations from direct links. Concerned about bias due to data that may be systematically missing with this approach, IGERT trainee Ashton Verdery worked with faculty in sociology and mathematics on an agent-based model to simulate the consequences of links that are missing due to mortality and migration for kin networks in a known case, Northeast Thailand. Simulating over 100 years, Verdery found substantial bias in the identification of extended kin ties, underestimating them by over 50%. He also demonstrated that this bias could be substantially reduced if surveys were to simply add questions about links to siblings. Verdery reported his results at the International Sunbelt Social Network for Social Network Analysis in March 2009.