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

Table 2

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	2.741	9.38	2.478	3.425	2.352	4.184	0.887	0.857	1.035
	10	2.854	9.608	2.632	4.753	2.417	4.695	0.812	0.894	0.908
	20	2.725	10.495	2.583	6.315	2.361	6.173	0.804	0.946	0.85
	40	2.814	16.952	2.645	12.214	2.432	11.902	0.821	0.981	0.837
10	5	3.012	4.841	2.801	2.716	2.553	2.643	0.852	0.919	0.927
	10	2.937	4.977	2.782	3.056	2.523	2.916	0.829	0.959	0.864
	20	2.958	7.656	2.841	6.01	2.576	5.813	0.831	1.011	0.822
	40	2.989	9.788	2.805	7.64	2.544	7.319	0.843	1.032	0.817
20	5	2.946	3.23	2.851	1.957	2.574	1.842	0.833	0.961	0.867
	10	3	3	2.799	2.203	2.524	2.087	0.834	1.008	0.828
	20	3.063	4.107	2.891	3.148	2.608	2.994	0.839	1.031	0.813
	40	2.995	5.178	2.842	4.182	2.551	3.934	0.857	1.06	0.808
40	5	3.052	2.199	2.869	1.518	2.594	1.427	0.831	1.001	0.83
	10	3.117	3.022	2.975	2.44	2.678	2.348	0.845	1.031	0.82
	20	3.033	2.871	2.87	2.24	2.57	2.129	0.858	1.053	0.815
	40	3.048	3.864	2.874	3.142	2.574	2.987	0.862	1.072	0.805