Gaming the Advice? Motivated Reasoning in Dynamic Voting Advice Applications

Voting Advice Applications (VAAs) are widely used tools intended to help citizens find parties or candidates that best match their policy preferences. So far, not much is known about how users interact with these tools. We argue that VAA design can fundamentally shape the cognitive and motivational processes underlying response behaviour. Building on theories of motivated reasoning, information effects and persuasion, we test whether giving users real-time information about party matching while responding induces them to strategically adjust their answers in order to preserve consistency with prior partisan preferences. We run an online experiment using a representative sample of eligible voters in Luxembourg. Participants are randomly assigned to either a standard VAA, which reveals users’ party matches only after questionnaire completion, or a dynamic VAA, which displays continuously updated party-match scores after each user’s response to the online questionnaire. Our core hypothesis is that in comparison to the standard, static VAA version, real-time feedback on party alignment activates directional motivated reasoning, leading users to strategically adapt their responses to maintain alignment with their preferred party. We further hypothesize that this will be especially the case for strong partisans and also explore the role of political sophistication as a potential moderator. We test these hypotheses on user behaviour (answer revisions, response times) and outcome (congruence between pre-existing party preferences and final VAA recommendations) measures. Our results speak to broader debates about motivated reasoning, preference measurement, and the role of VAAs as interactive informational tools.