In particular this concerns the way users can change the default test to apply for getting p-values (wilcoxon or permutation test) as well as the null hypothesis for the permutation test. Seb previously suggested the following:
Also given the prominence of the test type in the documentation perhaps make test_type an explicit argument rather than passing it via ellipsis?
As discussed in https://github.com/epiforecasts/scoringutils/pull/737, there are a few things that we could improve/make clearer related to the workflow of
get_pairwise_comparisons()
.In particular this concerns the way users can change the default test to apply for getting p-values (wilcoxon or permutation test) as well as the null hypothesis for the permutation test. Seb previously suggested the following: