Do voting advice applications have durable causal effects? Evidence from concurrent survey and field experiments

Abstract

Voting advice applications (VAAs) are used by millions of voters, yet their causal effects on political knowledge, preferences, and behavior remain unclear. We present new evidence from two experiments conducted during the 2024 UK General Election: (1) a panel survey experiment that provides access to participants’ VAA advice, achieves high compliance with treatment assignment, and allows us to track outcomes over time; and (2) a concurrent field experiment embedded in a real-world VAA that allows us to assess the ecological validity of our findings. We find that VAA usage leads to durable changes in voters’ perception of how close they are to parties on policy issues, shifts party preferences in line with the advice received, and affects the vote choices of undecided voters. These results suggest that VAAs play an important role in election campaigns. However, we find no evidence that VAAs increase electoral participation.