Empirical Analysis of Spatial Interdependence

Scott J. Cook, Jude C. Hays, and Robert J. Franzese


OVERVIEW: Social science research attempts to investigate, understand, and explain human behavior. Why do actors (e.g., people, groups, governments) behave as they do? What drives variation in social or historical phenomena over time and space? It is widely understood that these behaviors and outcomes do not occur in a social vacuum, but are instead conditional upon the context in which they take place. This includes location- and time-specific features of an actor's environment and the interactions amongst actors within these settings. As a result of shared environmental features or explicit interactions, the behaviors and outcomes of individual actors are rarely independent. This dependence complicates statistical inference in empirical analyses of social phenomena. More than just indicating the need for researchers to take some remedial action, understanding the source and nature of this dependence is itself substantively meaningful.

Distinct sources of correlation across behaviors and outcomes suggest fundamentally distinct theoretical understandings of these phenomena. Importantly, when this correlation reflects interdependence - that units affect and are affected by the choices, actions, and outcomes of one another – our theoretical and statistical models should reflect these interactions. Interdependence is theoretically and substantively central to most of the social sciences and, in a sense, interdependence is what makes the social sciences social. It spans the scope and substance of research in the social sciences. Our book highlights the theoretical importance of interdependence and, equally important, modeling spatial effects correctly such that we can draw accurate inferences about interdependence. While social science research suggests increasing awareness of the former, it has not always been paired with the latter. Often, we feel, this is because researchers are either unaware of the challenges of, or unfamiliar with the techniques to, estimate spatial models from which they can draw sound inferences about theories of interdependence. Our book aims to redress both of these potential impediments to broader use of spatial models in applied research.