Table 2.

Same as Table 1 but with sample size n = 100,000.

dim = 2					dim = 20
p	u	EMD	QF	Chi-square	p	u	EMD	QF	Chi-square
0	0	0.017	0.001	0.161	0	0	0.529	0.0032	0.165
0.001	1	0.015	0.001	0.165	0.001	1	0.525	0.0033	0.162
0.001	5	0.019	0.002	0.165	0.001	5	0.528	0.0031	0.165
0.001	10	0.026	0.002	0.16	0.001	10	0.528	0.0033	0.164
0.01	1	0.021	0.003	0.163	0.01	1	0.534	0.0034	0.168
0.01	5	0.055	0.008	0.173	0.01	5	0.532	0.0048	0.165
0.01	10	0.108	0.01	0.175	0.01	10	0.598	0.011	0.176
0.1	1	0.104	0.024	0.171	0.1	1	0.54	0.011	0.167
0.1	5	0.497	0.076	0.255	0.1	5	0.855	0.065	0.247
0.1	10	1.001	0.1	0.263	0.1	10	1.353	0.099	0.264
1	1	0.996	0.227	0.537	1	1	1.151	0.125	0.434
1	5	4.993	0.755	1.964	1	5	5.425	0.665	1.957
1	10	10	0.988	2	1	10	10.225	0.961	2

Only one simulation was performed in the null case p = 0 due to the computational burden of EMD. The results confirm the conclusions of Table 1.