VAAs and Their Impact on the Quality of Political Representation

Recent studies have shown strong indications that VAAs have an impact on the electoral decision-making of voters using them (Pianzola 2014, Vassil 2012). However, these studies focussed solely on the impact of VAAs on the voters’ party choice but did not address any questions related to a potential impact on the outcome of the representational processes as a whole. Our paper uses these studies as starting point and adds an expanded research focus by including the question, whether VAAs affect the outcome and the quality of the representational system in general. In order to do so we rely on two new and unique data sets: First, data from the Swiss VAA smartvote (N=3,205 or 84% response rate) from the 2015 elections, which provides comprehensive information on the positions of candidates and parties in both a multi-dimensional political space and a large number of specific and very detailed issues. Second, data from the Swiss Electoral Study (Selects). For the first time ever Selects included in the questionnaire for the 2015 elections not only the question, whether voters had used a VAA or not, but also integrated a number of issue questions used by VAAs in the 2015 election. Thus the Selects data provides a representative sample (N=6,500) of voters containing all the information of a complete national election study and in addition also comprehensive information on voters’ VAA use and voters’ positions on the questions the VAA had relied on. Do VAAs improve the quality of representational processes? This is the overall research question of our paper. All VAAs imply – at least implicitly – that voters should vote for those parties with the best match to their own values and issue preferences, and that representation works best, if the gap regarding the political positions of elected candidates and voters is as small as possible. Our data allows the conduct of an in-depth analysis, whether VAA users show a better match with the candidates they had voted for than voters not using a VAA. We will test the congruence on three levels: the left-right dimension, a multi-dimensional political space, and on the level of a number of specific policy issues. Thus we can control for the effect Thomassen (2012) called “the blind spot of representation”. He argued that political representation works well, but only as long as it is measured as voter-candidate/MP congruence on the left-right dimension. If measured as congruence on specific policy issues the picture would change dramatically. Our data also allows controlling for a large number of additional potentially interfering effects like the saliency of specific issues or the level of political interest and knowledge. By combining data from a VAA and a national election study, and by addressing a research question, which is crucial for VAA related research as well as it is for research on political representation (e.g. for discussions on correct voting or saliency theory), our paper will hopefully provide a further step forward heading for a closer link between these fields of research.