The effectiveness of interventions addressing conspiracy beliefs: a meta-analysis

Reducing conspiracy beliefs through effective interventions may help mitigate potential harmful consequences, such as vaccine hesitancy and prejudice. Therefore, a systematic literature search was conducted in Web of Science, PsycINFO, and Google Scholar for experiments testing interventions that could potentially reduce conspiracy beliefs. The present Bayesian three-level meta-analysis with 273 effect sizes from 56 samples (N = 27,996) provides the first comprehensive overview of factors affecting intervention effectiveness, including sample characteristics, intervention design and conspiracy belief features. Although the average intervention effect was small (g = 0.16, 95% CR [0.12, 0.20]), individual effect sizes exhibited substantial variability. Interventions adopting several evidence-based recommendations, such as fact-checking conspiracy claims, tailoring intervention content to specific beliefs and avoiding single-item measures of conspiracy beliefs, can potentially achieve effects exceeding g = 0.3. These findings advance theoretical models on the role of sender, message and recipient characteristics in combating harmful conspiracy beliefs. At the same time, extensive sensitivity analyses revealed that various bias sources might lead to under- or overestimation of certain effects. Furthermore, although study-level risk of bias scores (RoB 2 tool) were unrelated to intervention effectiveness, 82% of the available studies were at high risk of bias. The identified limitations and research gaps, such as the prevalence of underpowered studies and the scarcity of research on long-term intervention effects, provide clear directions for strengthening future intervention research.