Marcelo Escolar

The Local Voter: A Geographically Weighted Approach to Ecological Inference

expressed in the spatial structure of much ecological data. In this article we propose a Geographically Weighted Regression approach (GWR) for solving problems of spatial aggregation bias and spatial autocorrelation that affect all known methods of ecological inference. The estimation process is theoretically intuitive and computationally simple, showing that a wellspecified GWR approach to Goodman and King’s Ecological Inference methods may result in unbiased and consistent local estimates of ecological data that exhibit extreme spatial heterogeneity.

Ultimas imagenes antes del naufragio: las elecciones del 2001 en la Argentina

Resumen: Las elecciones de octubre del 2001 constituyen, sin duda, una coyuntura clave para interpretar el actual proceso de reestructuracion del sistema de partidos en la Argentina. Por primera vez en la historia, y sin el Peronismo proscrito, los dos partidos mayoritarios obtuvieron menos de la mitad de los votos validos emitidos y, ademas, el voto blanco, el voto nulo y el voto por terceras fuerzas acumularon el 56% de las preferencias ciudadanas. Uno de los elementos fundamentales para la interpretacion de este fenomeno es la estimacion correcta de la procedencia partidaria de los votos que lo construyeron. En este articulo se estiman las transferencias de votos producidas desde las fuerzas politicas mayoritarias de 1999 hacia el “voto protesta” y las terceras fuerzas en el 2001. Los resultados, obtenidos mediante la aplicacion de tres modelos distintos de inferencia ecologica (Goodman, EI y GWR-EI), muestran diferencias significativas entre los partidos politicos analizados. En este sentido, la mayor transferencia de votos Aliancistas hacia terceras fuerzas contrasta con las moderadas transferencias al voto blanco observadas desde el Peronismo o la importante transferencia de Accion por la Republica hacia el voto nulo. Este estudio de transferencias del voto 1999-2001 permite observar los realineamientos de preferencias de los votantes y explicar las significativas diferencias geografico-partidarias que acompanan la crisis y reestructuracion del sistema politico argentino.

Ecological Inference: Places and Relationships in Ecological Inference

One of the most salient but less studied features of ecological inference is the presence of spatial structure inducing aggregation bias in the observed data. This lack of attention is due to the fact that in most ecological inference models aggregation bias and spatial aggregation bias have been confounded into one and the same thing. However, provided that we know the location of the observable ecological units, there exists considerable more information about spatial aggregation bias than about most other non-spatial sources of bias. In this article we take advantage of a geographically weighted auto-regressive approach (GW-AR) to ecological inference that incorporates information about the underlying sources of spatial aggregation bias in ecological data. This spatial information can be then incorporated into most ecological inference methods although we will focus on spatial auto-regressive controls for the Goodman regression and King’s EI. In doing so, we will also shed light on the different performance of the standard Goodman and EI models in the presence of spatial effects (Anselin and Tam Cho, 2002; King, 2002; Calvo and Escolar, 2003) and their different local estimates (Herron and Shots, 2001; Adolph and King, 2002; Herron and Shotts, 2002). There are a number of different procedures that can be used to explore spatial aggregation bias in ecological data. Geographically Weighted Regression (GWR) provides a theoretically sound and computationally simple alternative within the classical framework. We also provide a distance weighted MCMC alternative in Apendix A, in the spirit of that presented by Haneuse and Wakefield in Chapter 13 of this volume.