Clarification on scale_test_by on bootstrap_ab

[CooperData]

I noticed one of the parameters in bootstrap_ab is scale_test_by. The docstring reads:

scale_test_by: The ratio between test and control population sizes. Use this if your test and control split is different from a 50/50 split. Defaults to 1.0.

However, looking at the implementation, it's using this ratio to multiply the whole dist. Is this value only needed when using stat_functions.sum? It's not clear to me why I would scale the values on my test sample when I am testing the mean on both samples.

test_ctrl_dist = compare_func(test_dist * scale_test_by, ctrl_dist)