Testing the mode effect hypothesis: A comparison of the response styles of traditional survey respondents and online (VAA) respondents

Despite their growing popularity and increased use of their data in political science, the issue of whether Voting Advice Applications (VAA) create data equivalent to those of traditional surveys remains understudied. Beyond producing datasets non-representative of national populations, VAAs differ from other modes of data-collection in that they involve self-selected individuals, provide immediate feedback on the basis of users’ responses and are answered in a self-completing capacity; all three aforementioned characteristics have been connected to socially desirable responding and fatigue-related effects on responses, such as satisficing or primacy effects. For this study data from EUvox2014, a VAA created for the 2014 European parliament elections, and data from the 2014 European Election Study (EES), collected through Computer Assisted Personal Interviews, are compared to examine differences among respondents from 21 EU member states. We use propensity score matching techniques to create matched for pairs of respondents based on a number of socio-demographic and political orientation variables. We then compare the two groups with a particular focus on overall responding style and party propensity to vote (PTV) scores.