Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published March 2014 | public
Journal Article

Party preferences in the digital age: The impact of voting advice applications


Election campaigns in modern democracies are increasingly affected by the rise of Internet-based information and communication technologies. Here, one of the most significant developments concerns the rapidly spreading 'voting advice applications' (VAAs). VAAs help online users to find their parties by matching a series of party positions with the positions taken up by the users themselves. This article analyses whether such tailor-made campaign information affects the party preferences of voters. It hypothesizes that the impact of the VAA-produced information is dependent on the 'representative deficit', a concept which captures the part of political preferences of a user that are not reflected by any party in the political system. The empirical analysis is conducted using a dataset stemming from users of the 'EU Profiler', a VAA produced for the 2009 European Parliamentary elections campaign. We find that less than one out of five users matches best with her initially preferred party. While a large majority of those users who are advised to vote for a better-fitting party are hardly affected by the VAA output, 8 percent of these do change their first preference to the party proposed by the VAA. Switching party preferences can be explained by the size of the representative deficit: the higher the deficit, the lower the probability that the initial party preference will be changed. This finding shows that users of VAAs are responsive to the match calculated by the online tool. They do not blindly follow the personalized suggestions but do so only when they are shown convincing levels of overlap between their views and the best-ranked party's positions.

Additional Information

© 2014 The Authors. Paper submitted 20 January 2013; accepted for publication 15 December 2013. Published online before print February 9, 2014. Funding: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. Acknowledgements: The idea for a pan-European Voting Advice Application originated during a Departmental Seminar session at the European University Institute in 2007. As Head of Department, Peter Mair chaired the session and from that moment became one of the leaders of the large-scale EU Profiler project. His intellectual energy and his dedication (he travelled to Antwerp and Brussels on two different occasions to help conceive and launch the tool, and as part of the Irish country team he worked on coding the positions of the Irish parties) were quintessential to the success of the project. 1t was clear that he would be the author of several contributions using EU Pro filer data. A first publication (Trechsel and Mair, 2011) presented the project, the data and used party-generated information. Our main goal, however, was to produce an article (or two) on the effects of the Pro filer on its users. Our main aim was to answer the question: Do users change their party preferences after being exposed to the EU Profiler? The online tool was designed explicitly to answer this question. The data generated became an invaluable and unique source of information. During the launch of the EU Profiler at Scotland House in Brussels, Peter took the floor telling the journalists present that the data gathered would constitute 'the Holy Grail of political science'. He exaggerated, but we had a good laugh, and even though the data may not live up to those expectations, it is exceptionally rich and innovative. From the outset, Mike Alvarez wanted to work with us on this project, and in the summer of 2010 he spent a week at the EUI, where the three of us began working on the very large and complex dataset. We had a first draft of the article when Peter passed away. Since that time, Ines Levin has helped Mike and I with analysis of the complex data - she had already helped us with estimation of the preliminary models when Peter was still alive and she is now a co-author. With the agreement of Peter 's wife, Dr. Karin Tilmans, we are honoured to have Peter as co-author of our jointly initiated research output.

Additional details

August 22, 2023
October 26, 2023