Attempting to measure the intensity of opposing feelings in elections: A polarization approach to Catalonia’s independence case

Abstract

This paper addresses the Catalonian political process for independence. Based on the outcome of the 2017 Catalonian Regional Election, it firstly innovates by building an independence-feeling indicator for each municipality. Subsequently, we resort to the concept of polarization to assess the level of inherent conflict. Besides, we evaluate to what extent polarization can be explained by different variables. Then, in the second part of the paper, we redo the analysis but, to handle the potential aggregation problem that arises with geographically reference data (MAUP problem), we compute for each municipality two (antiindependence-feeling and proindependence-feeling) indicators rather than one. In doing so, we avoid merging votes that offset each other and, by applying the polarization index, we not only test for (inter-) polarization but also (intra-) polarization (within municipalities). The results show important differences. More specifically, although polarization across municipalities is relatively small, there is a high level of polarization within them. Moreover, while population density, Catalan proficiency and place of birth are instrumental when explaining polarization if a single indicator is computed, in the case of two indicators unemployment rate, age and wealth also emerge as remarkable variables, while among the three previous ones the importance of population density decreases.