Center for Political Studies
Institute for Social Research
University of Michigan
P.O. Box 1248
Ann Arbor, MI
48106-1248

Voice: (734) 763-1348
Fax: (734) 764-3341

Title
Diagnosing, Modeling, Interpreting, and Leveraging Spatial Relationships in Time-Series-Cross-Section Data

PI
Robert Franzese, Jr. (Center for Political Studies, University of Michigan, and Political Science, University of Michigan

Direct Source
National Science Foundation

Abstract
Analysts recognize that observations in time-series-cross-section (TSCS) datasets will usually correlate across time and space. Their analyses typically reflect a view of these temporal and spatial dependencies from one of two perspectives. Some are concerned about drawing faulty causal inferences about theoretically chosen explanatory variables, but their interest in temporal and spatial dependence ends with this concern. They do not always see the need to model these relationships. They argue--sometimes incorrectly--that the only cost is reduced efficiency, and that as long as robust standard errors are used their inferences are sound. Others take a substantive interest in these relationships and attempt to model them directly.

In this project, we use Monte Carlo experiments to evaluate the performance of several simple and sophisticated estimators under three important types of spatial correlation. We develop a set of techniques and guidelines to help analysts diagnose spatial correlation and choose appropriate estimators based on their objectives.