Equivariance and Generalized Inference in Two-Sample Location-Scale Families

Table 6

Comparison with the bootstrap method (normal case with unequal sample sizes).

Sizes (𝑚,𝑛)	(𝛿,𝜎1,𝜎2)	Coverage proba.		Average width		St. error of width
		Bootstr	GTV	Bootstr	GTV	Bootstr	GTV
(5,10)	(0,2,1)	0.941	0.952	5.199	4.802	2.605	1.607
(5,100)	(0,2,1)	0.946	0.948	6.284	4.290	3.499	1.661
(50,100)	(0,2,1)	0.948	0.950	1.184	1.197	0.117	0.108
(10,5)	(0,2,1)	0.953	0.955	3.424	3.702	0.726	0.793
(100,5)	(0,2,1)	0.934	0.954	2.493	2.362	1.000	0.699
(100,50)	(0,2,1)	0.944	0.951	0.963	0.972	0.067	0.056
(5,10)	(0,200,2)	0.942	0.948	764.452	466.407	769.088	171.111
(5,100)	(0,200,2)	0.946	0.952	698.5501	470.437	504.280	172.287
(50,100)	(0,200,2)	0.953	0.950	111.983	113.287	12.110	11.434
(10,5)	(0,200,2)	0.945	0.952	297.40	278.228	78.240	65.770
(100,5)	(0,200,2)	0.942	0.951	78.867	79.232	6.198	5.604
(100,50)	(0,200,2)	0.949	0.950	78.666	79.242	6.155	5.601