Assessing the Robustness of Mediation Analysis Results Using Multiverse Analysis.

ABSTRACT: There is an increasing awareness that replication should become common practice in empirical studies. However, study results might fail to replicate for various reasons. The robustness of published study results can be assessed using the relatively new multiverse-analysis methodology, in which the robustness of the effect estimates against data analytical decisions is assessed. However, the uptake of multiverse analysis in empirical studies remains low, which might be due to the scarcity of guidance available on performing multiverse analysis. Researchers might experience difficulties in identifying data analytical decisions and in summarizing the large number of effect estimates yielded by a multiverse analysis. These difficulties are amplified when applying multiverse analysis to assess

SUBMITTER: Rijnhart JJM 

PROVIDER: S-EPMC9283158 | biostudies-literature | 2022 Jul

REPOSITORIES: biostudies-literature