- Altering W (simulation results)
- Investigating the role of W
- Choosing mu
- Results
- Iteratively correcting

To try and improve our estimate of \(\mu\) in problematic regions, such as high max z-scores in low power scenarios, we tried changing the prior standard deviation of the effect size from \(W=0.2\) to \(W=0.1\). We hoped that this would down-weight those values in problematic regions. Instead, we found that it made basically no difference in estimating \(\mu\).

The `.Wâ€™ methods are for \(W=0.1\).