Figure 8.
Median number of published studies across all simulation iterates (rows, numbers of studies in the underlying population before publication bias (5M
*); columns, distributions of study level random effects and true mean μ): , not selected on standard error;
, selected for small standard error;
, common effect;
, robust clustered;
, robust independent