Testing for heterogeneous effects in a case-control setting.
The aim here to discover an association that is beyond a mean difference
between all cases and all controls. Instead, the signal of interest here
is present in only a proportion of the cases. This test should be more
powerful than a t-test or Wilcoxon test in this heterogeneous setting.
Please cite the corresponding paper: Mezlini et al. (2020)
.