Table 9.

Root Mean Squared Error (MSE) and Bias for the estimation of the marginal survival probability S(1000) for the three groups defined by the number of nodes received

Generation		Estimation					Unadjusted		Weighted		Doubly Robust
PM	SM	PM	SM	n	Nodes	S(1000)	$\sqrt{\text{MSE}}$	Bias	$\sqrt{\text{MSE}}$	Bias	$\sqrt{\text{MSE}}$	Bias
2	3	3	4	1000	1-10	0.4108	0.0948	-0.0329	0.1144	0.0423	0.1175	0.0671
					11-20	0.2971	0.0283	-0.0044	0.0282	-0.0010	0.0258	0.0051
					3	0.5514	0.1056	0.0727	0.1250	-0.0866	0.1015	-0.0641
2	3	3	4	5000	1-10	0.4067	0.0498	-0.0312	0.0668	0.0483	0.0733	0.0589
					11-20	0.2941	0.0131	-0.0046	0.0124	-0.0010	0.0120	0.0039
					3	0.5461	0.0764	0.0687	0.1005	-0.0926	0.0823	-0.0751
2	3	3	4	10000	1-10	0.4069	0.0391	-0.0289	0.0575	0.0478	0.0643	0.0565
					11-20	0.2936	0.0099	-0.0045	0.0090	-0.0010	0.0088	0.0034
					3	0.5473	0.0708	0.0666	0.1008	-0.0969	0.0840	-0.0804

Both the propensity score model and the survival model used in the estimation process differ from the corresponding models used for data generation. Results are based on 3000 simulated samples. PM stands for Propensity Score Model, and SM stands for Survival Model. Please refer to the text for the description of the models