Spatial Effects and Ecological Inference

This paper examines the role of spatial effects in ecologicalinference. Both formally and through simulation experiments,we consider the problems associated with ecological inferenceand cross-level inference methods in the presence of increasingdegrees of spatial autocorrelation. Past assessments of spatialautocorrelation in aggregate data analysis focused on unidimensional,one-directional processes that are not representative of thefull complexities caused by spatial autocorrelation. Our analysisis more complete and representative of true forms of spatialautocorrelation and pays particular attention to the specificationof spatial autocorrelation in models with random coefficientvariation. Our assessment focuses on the effects of this specificationon the bias and precision of parameter estimates.