It was not at all intuitively obvious how one could carry out even a simple resampling inference for the difference in two means. As best I could figure out, it appears one has to opt for “Difference in Proportions” and then later in the process select “Difference in means” instead of “Difference in counts”? The output is similarly difficult to interpret as graph titles and labels are not clear and tick labels of the graph are not “pretty” and often overwrite so as to make reading the graphs impossible.