Correcting social norm misperceptions to counter online vaccine misinformation during pandemics

Abstract Online vaccine misinformation circulating on social media undermines public health efforts, yet factual debunking alone has shown limited success in improving vaccination intentions. We tested whether combining descriptive social norm information with factual debunking improves the correction of online vaccine misinformation, alters its mechanism of influence, and varies in effectiveness by prior COVID-19 vaccination status. In an online experiment, 731 UK adults were randomly assigned to view online vaccine misinformation about a hypothetical pandemic (Covid-66) under one of three conditions: debunked combined with descriptive norm messaging, debunked alone, or not corrected at all. Vaccine intentions and perceived peer vaccination norms were measured before and after exposure. Uncorrected online vaccine misinformation significantly reduced vaccine intentions by 2.7 percentage points (p < 0.001) and induced substantial norm misperception, with participants underestimating peer vaccine support by 8.04 percentage points (p < 0.001). Both correction conditions increased vaccine intentions relative to no correction, but the combined intervention did not produce significantly greater change than debunking alone (p = .681); our primary hypothesis of an incremental benefit from adding descriptive norm information to debunking was not supported on this average-effect outcome. Exploratory subgroup analyses indicated possible differential effects by COVID-19 vaccination status. Among participants who were unvaccinated or had received only one dose, the combined intervention produced significantly higher post-intervention intentions than exposure to misinformation alone, whereas debunking alone reached only borderline significance (p = .05). Among participants who had received two doses without a booster, the combined intervention produced significantly higher intentions than both debunking alone and no correction. Among boosted participants, intentions were high and comparable across conditions. Across the five posts, which targeted vaccine safety, vaccine efficacy, and institutional trust, exposure reduced vaccine intentions largely by leading people to underestimate how much others support vaccination (pluralistic ignorance): the indirect effect through norm misperception was b = − 6.80 (95% CI [− 10.32, − 3.73]), with no direct effect detectable. Both correction strategies reduced this underestimation of peer support, but to very different extents. The combined intervention almost completely corrected the misperception, and its effect on intentions was carried by that pathway with no residual direct effect detected. Debunking alone only partially corrected the misperception and showed a smaller indirect effect alongside a direct effect at the conventional threshold, suggesting it also operates through mechanisms not measured here. Taken together, these findings suggest that online vaccine misinformation during pandemics can reduce vaccine intentions partly by distorting perceptions of what peers do, and that debunking partially corrects this misperception. Adding descriptive norm information more fully corrects the norm misperception but does not, on average, produce a detectable additional benefit in intention change relative to debunking alone. Exploratory analyses suggest the combined intervention may be more helpful than debunking alone among less-vaccinated participants, a pattern consistent with norm interventions being most effective where room for attitude change is greatest, though these warrants preregistered confirmatory testing.