CAREER: Social and Economic Consequences of Information Diffusion in Networks

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

This project is a three-pronged program of research and education focused on information in networks - its distribution, diffusion, value, and consequences for social and economic outcomes. First, this project will map the flow of information in real organizations over time and will combine economic theories of production with social network theory to estimate the effects of information diffusion on the productivity of individuals and teams. This work will advance economic theory concerning information worker production and productivity; sociological theory concerning the movement of information in social networks; technical methods for analyzing information content in digital networks while preserving privacy; and statistical methods to identify causal relationships between network structure and productivity. Second, the research will utilize data on online social networks to identify estimates of peer influence in product adoption and demand, and it will develop and apply new statistical techniques to separate influence from selection, homophily and confounding factors to optimize targeted peer-to-peer information diffusion. Third, the research will develop a general class of network-based models of the diffusion of behavior change in social networks, and parameterize and validate these models using empirical data on social relationships and health related behaviors from five massive networked datasets. The central goals of the project are: 1) to estimate and enhance the productivity of information workers, 2) to model, measure and improve peer-to-peer viral marketing and demand estimation, and 3) to improve health-related prevention and intervention strategies. The proposed activities are designed to have broader impact for science, education and society. Our understanding of information worker productivity is a cornerstone of our future economic growth and thus our prosperity and welfare. The project also aims to improve intervention and prevention strategies for behaviors and outcomes such as disease prevention, obesity, and drug use. Developing reliable, scalable statistical methods for establishing causality in the diffusion of disease, obesity, smoking and delinquency can change social policy to ensure health interventions are appropriately targeted. Development of technical infrastructures to deal with large dynamic networked datasets, privacy preserving text analysis, and statistical methods for estimating the impacts of information diffusion in networks have implications for telecommunications, epidemiology, medicine, social policy and any industry or social sphere characterized by network externalities and local network effects. Educationally, the project will utilize the research program in classes to highlight the importance of networks and information for business strategy, society and health. Workshops will develop a new collaborative research community across academia and industry around topics related to information diffusion in networks.