Empirical Power Functions for Nonparametric Two-Sample Tests for Small Samples