Dasymetric distribution of votes in a dense city

Abstract

[EN] A large proportion of electoral analyses using geography are performed on a small area basis, such as polling units. Unfortunately, polling units are frequently redrawn, provoking breaks in their data series. Previous electoral results play a key role in many analyses. They are used by political party workers and journalists to present quick assessments of outcomes, by political scientists and electoral geographers to perform detailed scrutinizes and by pollsters and forecasters to anticipate electoral results. In this paper, we study to what extent more complex geographical approaches (based on a proper location of electors on the territory using dasymetric techniques) are of value in comparison to simple methods (like areal weighting) for the problem of reallocating votes in a large, dense city. Barcelona is such a city and, having recently redrawn the boundaries of its census sections, it is an ideal candidate for further scrutiny. Although previous studies show the approaches based on dasymetric techniques outperforming simpler solutions for interpolating census figures, our results show that improvements in the process of reallocating votes are marginal. This brings into question the extra effort that entails introducing ancillary sources of information in a dense urban area for this kind of data. Additional research is required to know whether and when these results are extendable. (C) 2017 Elsevier Ltd. All rights reserved.