One model to rule them all? Choosing two-dimensional spaces for European political landscapes with VAA data

Thomeczek, Jan Philipp; Zhirnov, Andrei; Scotto di Vettimo, Michele; Buma, Roan; Cristea, Vladimir; Dolle, Maximilian; Kersting, Norbert; Krouwel, André; Lange, Sven; Lopez Ortega, Alberto; Thompson, Jack; Antonucci, Lorenza; Di Stefano, Roberta; Ruggeri, Francsesco

Abstract

This article investigates the effectiveness of three different two-dimensional models in representing political competition using user data from the EU and I Voting Advice Application for the EU elections from 2009 to 2019. We test different ways of aggregating economic, cultural, and European integration issues to determine which configuration best captures voter preferences across all member states. Our findings show that combining economic and cultural issues into a left-right “super-dimension” alongside a separate dimension for European integration generally offers the most accurate representation across time and space. However, regional variations reveal that this model performs better in Western Europe compared to Central/Eastern Europe, where other models may fit better. To address these regional disparities, we recommend implementing more rigorous pre-testing of common items and considering country-specific or regional adaptations in scaling. These measures could enhance the relevance and precision of VAAs, ensuring they effectively reflect the political landscapes of diverse European regions.