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

Table 16

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	2.186	1.524	0.229	0.296	2.186	1.524	1.023	1.095	1.208
	10	2.192	2.151	0.224	0.396	2.192	2.151	1.011	1.076	1.125
	20	2.203	3.536	0.232	0.69	2.203	3.536	1.074	1.065	1.103
	40	2.21	6.428	0.232	1.173	2.21	6.428	1.051	1.06	1.086
10	5	2.307	1.207	0.109	0.154	2.307	1.207	1.081	1.094	1.08
	10	2.322	1.564	0.105	0.182	2.322	1.564	1.068	1.096	1.059
	20	2.316	2.34	0.106	0.25	2.316	2.34	1.052	1.098	1.049
	40	2.318	3.892	0.105	0.441	2.318	3.892	1.046	1.095	1.047
20	5	2.378	1.048	0.032	0.104	2.378	1.048	1.024	1.094	1.028
	10	2.363	1.229	0.065	0.11	2.363	1.229	1.027	1.096	1.028
	20	2.362	1.626	0.036	0.145	2.362	1.626	1.039	1.099	1.028
	40	2.377	2.462	0.048	0.204	2.377	2.462	1.028	1.101	1.024
40	5	2.387	0.932	0.026	0.092	2.387	0.932	1.011	1.092	1.017
	10	2.394	1.032	0.009	0.091	1.05	-0.002	1.008	1.089	1.017
	20	2.39	1.238	0.03	0.108	1.255	0.019	1.007	1.092	1.014
	40	2.404	1.669	0.02	0.142	1.693	0.01	1.011	1.096	1.014