SOCIAL NETWORKS, AGENT-BASED MODELS, AND DATA SCIENCE: FROM INDIVIDUAL DECISIONS AND SOCIAL INTERACTIONS TO AGGREGATE OUTCOMES

Understanding how the behaviors of individual people aggregate to produce socialpatterns is the foundational question in the social sciences. In the same way it is often hard to understand how the driving behavior of individuals leads to traffic jams, a wide variety of social phenomena are not well understood in terms of the actions of the people involved. In this dissertation I use the methods of modern computational social science, including agent-based models (ABM), social network analysis (SNA), and data science, to study how human communication leads to cross-border migration flows, how corruption arises via social influence, and how retention levels in higher education result from social interactions. In each of these three cases I focus on the decision-processes people engage in and then use computing tools to assess how myriad decisions scale to the aggregate level. Each case is different, but each has the common theme of micro motives leading to macro-outcomes. This is achieved by modeling patterns of behavior from the real world and seeing what emerges in the model, and how it compares to patterns of behavior observed in the real-world. This provides a way to promote interdisciplinary research and work towards more comprehensive and tested solutions to problems in the social sciences. The three essays provide three different examples of how this can be done, particularly with a lens of understanding the effects of external influence on decision-making and behavior.

Author

• Stine, Amira Y Al-Khulaidy

• Agent-based Models
• Computational Social Science
• Data Science
• Education
• Migration
• Social Networks

Identifier

hdl:1920/13823

OAI identifier oai:identifier

oai:MARS:1920/13823