The Impact of Voting Advice Applications on Electoral Preferences: A Field Experiment in the 2014 Quebec Election

Abstract

Voting Advice Applications (VAAs) provide nonpartisan and interactive information during election campaigns, and match the policy demand of users with the political supply of parties. But do VAAs influence citizens’ electoral decisions? Do they help the undecided to form a political preference, and lead the decided to change their vote choice? This article reports a randomized field experiment to evaluate the effect of the Vote Compass, a VAA used in Quebec, on users’ electoral preferences. Given Quebec's multidimensional political space, VAAs have the potential to assist citizens in making a complex electoral decision. The results show that the VAA's users are more likely to form an electoral preference, but this is only the case in the short term among 30‐year‐olds, the higher educated, and more politically interested users. At the same time, using a VAA mostly impacts party preferences of voters in the short term: the politically uninterested and older users are more likely to change their party preferences, whereas 30‐year‐olds and politically interested users are less likely to change their initial party preference. This study provides evidence that the VAA cognitively engaged users, but not that it leads to vote switching on Election Day.

Does receiving advice from Voter Advice Applications (VAAs) affect public opinion in deeply divided societies? Evidence from a field experiment in Northern Ireland

“…It seems that the large majority of citizens are unwilling to follow last-minute suggestions, even if they come from trustworthy sources and even if they are formulated in unequivocal form ' (2016: 1012-1013). Maheo (2016) states that 'receiving non-confirmatory advice (p. 406) did not significantly push users to change their party preferences (compared to other users who received confirmatory advice. See also Pianzola (2014) for similar findings.…”

“…Enyedi's results 'lend support to those studies that have found minimal impact in terms of direct preference change … Therefore, expectations concerning VAA-induced change need to be modest. It seems that the large majority of citizens are unwilling to follow last-minute suggestions, even if they come from trustworthy sources and even if they are formulated in unequivocal form' (2016: 1012-3) Maheo (2016). states that 'receiving nonconfirmatory advice did not significantly push user to change their party preferences (compared to other users who received confirmatory advice'.…”

“…Although several studies have concluded that VAAs have effects on party preference, the methodology is typically non-experimental and hence limited in terms of isolating the causal effect of VAA use on party support (Kamoen et al, 2015;Ladner et al, 2012;Walgrave et al, 2008;Wall et al, 2014). 1 Two exceptions include experimental work by Enyedi (2016) and Maheo (2016) which reveal minimal effects. 2 This means that applying VAAs to deeply divided places where politics is particularly crystallized may be regarded as an especially tough test for VAA effects.…”

Research on Voting Advice Applications: State of the Art and Future Directions

“…In her article in this issue, Mahéo () speaks to a consolidated subfield of VAA research—the effect of these tools on users’ voting intentions. She tackles one of the crucial methodological issues in this strand of literature, namely selection bias, through a full‐fledged experimental design in the context of the 2014 Quebec Elections.…”

Mapping the Austrian political spectrum with the help of VAAs

“…For example, more than 4m people used the StemWijzer before the 2012 elections in the Netherlands and more than 15m the Wahl-O-Mat ahead of the German national election in 2017. Research on VAAs has mostly dealt with their effects on political knowledge (Schultze 2014), issue voting (Ladner 2016), voter turnout (Fivaz and Nadig 2010;Marschall and Schultze 2012), electoral preferences (Mahéo 2016;Walgrave, van Aelst, and Nuytemans 2008;Kleinnijenhuis et al 2017) as well as with the impact of design choices on the recommendations given to voters (Germann et al 2015;Rosema and Louwerse 2016;Agathokleous and Tsapatsoulis 2016;Katakis et al 2014).…”