Mapping parties in a multidimensional European political space: A comparative study of the EUvox and euandi party position data sets

Abstract

This study compares the spatial positioning of over 200 political parties across 28 European Union (EU) member states in two cross-national voting advice applications (VAAs) developed for the 2014 European elections: EUvox and euandi. We find that the two VAAs show highly similar results in terms of party positioning on the cultural liberal-conservative and pro-anti EU dimensions, while economic left–right placements converge less, especially concerning right-wing parties. Our analyses reveal that the higher overlap on the cultural and EU dimensions is a result, at least partially, of the inclusion of similar items used to measure these concepts, while most of the systematic divergence between the two VAAs in left–right placements stems from problematic issue-statements used in the dimensional calculations. We demonstrate how certain items can cause bias in the placements of specific party families by (1) not aligning with other statements that measure the same latent construct; (2) tapping into other latent constructs, in addition to the one they are supposed to measure; and (3) not inducing sufficient polarization between parties.

Beyond the U-curve: Citizen preferences on European integration in multidimensional political space

“…Some scholars have considered whether the economic left-right, social left-right and EU dimensions are separate from each other (in particular, Bakker et al, 2012) and have found supportive evidence for this conjecture using data on party positions. In fact, most of the theoretical reasoning and the empirical analyses leading to the results summarized above are based on studying political party positions (Bakker et al, 2012; Gabel and Hix, 2002; König et al, 2017; Reiljan et al, 2019). But political space might look quite different when we consider the attitudes and positions of citizens rather than parties.…”

Assessing the impact of party identification in transnational election campaigns: Evidence from the 2019 European Parliament election

“…We can therefore assume that, if PID is characterised by ideology (Fiorina 1981), it should be feasible to translate the nationally formed party identification into a European party environment. Pre-requisites being met for this adaptation process at the micro-sociological level, however, are that at the macro-sociological level the underlying axes represent comparable cleavages (Lipset and Rokkan 1967) and span an equivalent coordinate system of political parties within the EU (Reiljan, Kutiyski and Krouwel 2019;McElroy and Benoit 2007;Gabel and Hix 2002). If we further relate this empirical expectation to consistency-theoretical approaches (for example Heider 1958;Festinger 1962), we can assume that in a TV debate those contents are perceived by the recipients that correspond to existing ideological attitudes in order to avoid cognitive dissonances.…”

Longitudinal dataset of political issue-positions of 411 parties across 28 European countries (2009–2019) from voting advice applications EU profiler and euandi

“…the statements must address the political issues that are important in the respective electoral context. Composing a list of divisive and salient statements in pan-European context is undoubtedly a challenging task [8] . A large number of sources – such as opinion polls, earlier party manifestos, information from experts, academics and journalists – has been utilized to create and revise this list of statements [2] .…”

“…With very few exceptions, the issue-statements in the dataset were attached either to economic left-right, cultural liberal-conservative or pro-/anti-EU dimension. Factor analyses conducted after the data collection waves have confirmed the described three-dimensional structure of the data and the a priori decisions regarding which statement should be assigned to which dimension have in most cases proven to be valid [8] [9] [10] . The three-dimensional political space visualization that was displayed to the euandi users in 2014 is shown on Fig.…”