Skip to main content

View full-text article in PMC

. 2013 Sep 11;15(4):1232–1241. doi: 10.1208/s12248-013-9526-y

Table I.

Relative Bias (in Per Cent) in Population Parameter Estimates for the Different Methods When Data Were MCAR, MAR and MNAR

	MCAR				MAR				MNAR
	CL_male	CL_female	BSV	ResErr	CL_male	CL_female	BSV	ResErr	CL_male	CL_female	BSV	ResErr
ALL	−0.32	−0.064	−1.7	0.82	−0.32	−0.064	−1.7	0.82	−0.32	−0.064	−1.7	0.82
EST	−0.37	−0.25	−1.6	0.82	−0.65	0.020	0.22	0.82	−0.41	−0.18	−1.4	0.82
MOD	−0.40	−0.10	−1.2	0.82	−0.88	−0.055	1.1	0.82	5.8	10	6.9	0.82
MI	−0.26	−0.054	−1.7	0.82	−0.69	0.14	0.75	0.82	5.2	6.2	−3.7	0.82
SI_WT	−11	11	69	0.83	−11	10	69	0.83	−3.1	32	87	0.83
SI_mode	−16	0.24	78	0.82	−14	3.2	76	0.83	−0.90	42	110	0.83
CC	−0.14	0.022	−2.0	1.3	−0.27	0.18	−3.6	0.60	−0.90	−0.051	−2.9	1.4

The between-subject variability (BSV) and the residual error (ResErr) were estimated as variances