The Influence of Voting Advice Applications on Preferences, Loyalties and Turnout: An Experimental Study Zsolt Enyedi Central European University Voting Advice Applications (VAAs) are increasingly popular, yet little is known about their impact. This article investigates their influence on party choice, party loyalty and electoral participation, relying on a complex experiment conducted before and after the 2010 Hungarian election. Participants were directed to two VAAs, some received advice from one and some from both, while the control group visited none. According to subjective recollections, 7 per cent changed their vote intentions, but according to the panel study the VAAs were unable to direct users to specific parties. Sheer exposure to the advice did not have mobilising or demobilising effects either, but preference- confirming outputs increased party loyalty while preference-disconfirming recommendations decreased it, and double exposure amplified further the impact of the VAAs. Converging advice from two different sources increased the rate of electoral participation, but more by provoking, rather than by persuading, the users. Keywords: voting advice applications; party choice; electoral participation; experiment; political parties Voting Advice Applications (VAAs) – that is, websites that identify the user’s proximity to candidates and parties based on a set of political questions – are becoming increasingly popular across Europe and have also reached North America, the Arab World and Latin America (Mendez, 2012; Wall et al., 2012). They aim to increase political knowledge, self-awareness and interest in elections and foster rational decision making. Apart from these goals, they have normatively neutral functions such as entertainment. Potentially, they can also manipulate. The more citizens use these tools, the more important it becomes for the public to know what goes on within the ‘black box’ of VAAs and what influence they have on the attitudes and behaviour of citizens. Are users stimulated to participate at elections and do they listen to the offered advice? The experiment presented below indicates that simple exposure to a VAA does not affect electoral turnout, and the number of VAA users who abandon their favourite party in order to follow the recommendations is very small. But VAAs do seem to have the power to consolidate or undermine party preferences. Those who received advice that was in line with their preferences were particularly likely to stay loyal to their party and if the preference-confirming recommendation came from two VAAs, then the level of party- switchers dropped further. Recommendations that contradicted a user’s original prefer- ences increased the number of those who abandoned their party. Finally, identical recommendations coming from the two VAAs led to a particularly high level of electoral participation – at least compared to contradictory outputs. The results indicate that under certain conditions VAAs may influence electoral behaviour but not primarily through channeling voters from one party to another. doi: 10.1111/1467-9248.12213 POLITICAL STUDIES: 2015 © 2015 The Author. Political Studies © 2015 Political Studies Association