The social influence of the corrections of vaccine misinformation on social media.

This study examines the impact of social versus algorithmic corrections of vaccine misinformation on social media, on the perceptions of social norms around vaccination and vaccine intentions during a hypothetical pandemic. In an online randomised study with 720 participants, we assessed whether user-generated or algorithmically generated corrections influenced perceived social norms measured as beliefs about other's vaccination intentions (empirical expectations about vaccination), perception about the social appropriateness of vaccine refusal (normative expectations), and beliefs about others' perceptions of vaccine safety (second-order normative beliefs), and own vaccine intentions. User-generated corrections significantly increased the perceptions that refusing vaccination is socially very inappropriate and increased the perception that consensus is in support of vaccine safety. Algorithmic corrections did not influence social norm perceptions. Interestingly, neither social nor algorithmic corrections significantly altered empirical expectations about others' vaccination intentions. Both corrections also helped maintain participants' high initial vaccine intentions with algorithmic correction having stronger effects, in contrast to a significant decline in intentions observed in the control group exposed to misinformation without corrections. Algorithmic corrections from credible health agencies were effective in sustaining vaccine intentions, while user-generated corrections were more influential in improving perceptions of social norms. The findings suggests that different correction types on social media may influence distinct determinants of vaccination behaviour during a pandemic: user-generated corrections reshape perceived social norms of vaccination, while algorithmic corrections, citing credible sources, may better sustain high personal vaccine intentions.