Development of New Robust Optimal Score Function for the Weibull Distributed Error Term in Multilevel Models

Table 14

Bias and MSE of using Weibull, Wilcoxon, and REML methods when  = 6..

		REML		Weibull		Wilcoxon		Precision
N2	N1	Bias	MSE	Bias	MSE	Bias	MSE	Weib/REML	Weib/Wil	Wil/REML
5	5	2.4	0.772	0.602	0.726	2.239	0.755	0.978	0.839	1.166
	10	2.553	1.263	0.214	0.138	3.129	1.543	1.022	1.089	1.028
	20	1.752	2.34	0.245	0.3	2.019	3.249	1.088	1.003	1.084
	40	2.545	9.357	0.798	5.724	3.11	9.185	0.982	0.867	1.033
10	5	1.473	0.399	1.303	2.546	1.348	0.701	1.057	0.887	1.98
	10	2.151	1.254	-0.362	0.262	2.205	1.155	0.921	1.094	0.771
	20	1.791	1.317	0.018	0.001	2.847	1.833	1.092	1.742	0.799
	40	2.144	3.688	-0.06	0.018	2.461	3.592	0.974	0.924	1.054
20	5	1.765	0.903	0.02	0.001	2.034	0.685	0.759	0.794	0.956
	10	1.584	0.928	0.092	0.013	1.577	0.875	0.943	1.096	0.86
	20	2.141	1.046	-0.353	0.25	2.334	1.182	1.030	1.063	0.971
	40	2.335	2.362	-0.06	0.011	2.601	2.925	1.038	1.091	1.04
40	5	1.907	0.947	-0.501	0.282	2.455	0.854	0.902	1.025	0.801
	10	2.11	0.841	0.296	0.109	2.476	0.924	1.099	1.054	0.952
	20	2.222	1.226	0.881	1.165	2.524	1.379	1.025	1.023	1.002
	40	2.321	1.615	0.086	0.015	2.615	1.758	1.089	1.079	1.009