This paper suggests an application of "P-P plots" for comparing two noncomparable variables with respect to the extent to which their distributions differ in two groups. The comparison is symmetric with respect to the variables and the groups. It reflects the differences between the groups over the full range of values of each variable.