Admissible ways of merging p-values under arbitrary dependence

Methods of merging several p-values into a single p-value are important in their own right and widely used multiple hypothesis testing. This paper is the first to systematically study admissibility (in Wald’s sense) p-merging functions domination structure, without any information on dependence structure input p-values. As technical tool, we use notion e-values, which alternatives recently promoted by authors. We obtain results representation admissible via e-values (in)admissibility existing functions. By introducing new functions, show that some classic methods can be strictly improved enhance power compromising validity under arbitrary dependence.