Preaching to the converted Misinformation and voter preferences in election campaigns

Politicians frequently sponsor misinformation during election campaigns, but its effectiveness in shifting voters' policy preferences and beliefs remains unclear. We argue that the efficacy of campaign misinformation depends on whether it latches onto partisan or non-partisan identities. Misinformation that primes non-partisan but politically relevant social identities could appeal to voters sharing these social identities, potentially moving supporters and non-supporters closer to the issue position of the sponsor of misinformation. If, on the other hand, misinformation is processed along partisan lines, it appeals only to existing supporters. Our empirical analysis is based on a pre-registered vignette experiment embedded in a representative post-election survey in India. Our design mimics the opposition's use of campaign misinformation to polarize voters' preferences and beliefs on issues of religion. We find that misinformation was only partly effective. While some co-partisans increased their support of the policy position advocated by the misinformer, messages failed to persuade non-supporters, and were easily corrected among co-partisans. We demonstrate the broader relevance of these results with a replication of campaign misinformation in the United States.