Social-Behavioral Modeling for Complex Systems

14 Combining Data‐Driven and Theory‐Driven Models for Causality Analysis in Sociocultural Systems

Amy Sliva¹, Scott Neal Reilly¹, David Blumstein¹ and Glenn Pierce²

¹ Charles River Analytics, Cambridge, MA, 02138, USA

² School of Criminology and Criminal Justice, Northeastern University, Boston, MA, 02115, USA

Introduction

One of the key aspects of nearly all social science research questions is to validate or discover an explanation for some type of causal relationship. These causal explanations provide a greater understanding of complex social and political dynamics, with the long‐term goal of influencing policy interventions or producing theories that can facilitate better decision‐making by policy‐makers. The growing accessibility of digital information promises many exciting research opportunities for expanding the breadth and depth of causal analysis. For example, studies of the relationship between conflict and natural resource constraints can pull economic data from any number of online sources (e.g. World Bank, World Trade Organization), records of violent incidents from event databases, political sentiments posted in online social media, in‐depth interviews from refugees provided through social science data exchanges or aid organizations, and even climate and water quality trends from environmental biologists, all with a few clicks of a mouse.

However, exactly how to leverage this big data environment remains a challenge for social scientists, who typically ...