Support for Censorship: A Multilevel Meta-Analysis of the Third-Person Effect

However robust the conceptual and operational definitions of the perceptual component of the third-person effect are, evidence for claims on the behavioral dimension of the theory (i.e., support for restrictions of undesirable messages) is mixed. Are variations in research findings real or simply methodological artifacts? Through a meta-analysis of 35 empirical studies, we addressed this and other major issues with a rigorous multilevel modeling approach. We found that the third-person effect size was on the whole rather weak. Analyses showed that country of study plays an important role in moderating the effect. Implications for theory development were discussed.