FIGURE 4.

Illustration of the AnCred with Bayes factors procedure using the findings from the meta‐analysis on the association of COVID‐19 mortality and corticosteroids. The left plot shows the Bayes factor ${\mathrm{BF}}_{01}$ as a function of the relative variance $g={\tau }^2/{\sigma }^2$ of the sceptical prior. The result from the RECOVERY trial is challenged with a sceptical prior such that ${\mathrm{BF}}_{01}=1/10,$ for the other trials such a prior does not exist. The right plot shows the Bayes factor ${}_{01}$ as a function of the relative mean $f=\mu /\widehat{\theta }$ of the advocacy prior where the coefficient of variation from the prior is fixed to $\mathrm{CV}=\tau /\mu =1/z\left(\gamma =1/3\right)=0.67,$ where $z\left(\gamma \right)$ is given in (20). The RECOVERY and the CAPE COVID findings are challenged such that ${\mathrm{BF}}_{01}=1/3$, for the other trials such a prior does not exist [Colour figure can be viewed at wileyonlinelibrary.com]